DeepCC: a novel deep learning-based framework for cancer molecular subtype classification

A review on trends in development and translation of omics signatures in cancer

The field of cancer genomics and transcriptomics has evolved from targeted profiling to swift sequencing of individual tumor genome and transcriptome. The steady growth in genome, epigenome, and transcriptome datasets on a genome-wide scale has significantly increased our capability in capturing signatures that represent both the intrinsic and extrinsic biological features of tumors. These biological differences can help in precise molecular subtyping of cancer, predicting tumor progression, metastatic potential, and resistance to therapeutic agents. In this review, we summarized the current development of genomic, methylomic, transcriptomic, proteomic and metabolic signatures in the field of cancer research and highlighted their potentials in clinical applications to improve diagnosis, prognosis, and treatment decision in cancer patients.

Artificial intelligence applied to 'omics data in liver disease: towards a personalised approach for diagnosis, prognosis and treatment

Advancements in omics technologies and artificial intelligence (AI) methodologies are fuelling our progress towards personalised diagnosis, prognosis and treatment strategies in hepatology. This review provides a comprehensive overview of the current landscape of AI methods used for analysis of omics data in liver diseases. We present an overview of the prevalence of different omics levels across various liver diseases, as well as categorise the AI methodology used across the studies. Specifically, we highlight the predominance of transcriptomic and genomic profiling and the relatively sparse exploration of other levels such as the proteome and methylome, which represent untapped potential for novel insights. Publicly available database initiatives such as The Cancer Genome Atlas and The International Cancer Genome Consortium have paved the way for advancements in the diagnosis and treatment of hepatocellular carcinoma. However, the same availability of large omics datasets remains limited for other liver diseases. Furthermore, the application of sophisticated AI methods to handle the complexities of multiomics datasets requires substantial data to train and validate the models and faces challenges in achieving bias-free results with clinical utility. Strategies to address the paucity of data and capitalise on opportunities are discussed. Given the substantial global burden of chronic liver diseases, it is imperative that multicentre collaborations be established to generate large-scale omics data for early disease recognition and intervention. Exploring advanced AI methods is also necessary to maximise the potential of these datasets and improve early detection and personalised treatment strategies.

Evolutionary screening of precision oncology biomarkers and its applications in prognostic model construction

Biomarker screening is critical for precision oncology. However, one of the main challenges in precision oncology is that the screened biomarkers often fail to achieve the expected clinical effects and are rarely approved by regulatory authorities. Considering the close association between cancer pathogenesis and the evolutionary events of organisms, we first explored the evolutionary feature underlying clinically approved biomarkers, and two evolutionary features of approved biomarkers (Ohnologs and specific evolutionary stages of genes) were identified. Subsequently, we utilized evolutionary features for screening potential prognostic biomarkers in four common cancers: head and neck squamous cell carcinoma, liver hepatocellular carcinoma, lung adenocarcinoma, and lung squamous cell carcinoma. Finally, we constructed an evolution-strengthened prognostic model (ESPM) for cancers. These models can predict cancer patients' survival time across different cancer cohorts effectively and perform better than conventional models. In summary, our study highlights the application potentials of evolutionary information in precision oncology biomarker screening.

Overexpression of TSPAN8 in consensus molecular subtype 3 colorectal cancer

BACKGROUND

Recently, consensus molecular subtypes (CMSs) have been proposed as a robust transcriptome-based classification system for colorectal cancer (CRC). Tetraspanins (TSPANs) are transmembrane proteins. They have been associated with the development of numerous malignancies, including CRC, through their role as "master organizers" for multi-molecular membrane complexes. No previous study has investigated the correlation between TSPANs and CMS classification. Herein, we investigated the expression of TSPANs in patient-derived primary CRC tissues and their CMS classifications.

METHODS

RNA samples were derived from primary CRC tissues (n = 100 patients diagnosed with colorectal adenocarcinoma) and subjected to RNA sequencing for transcriptome-based CMS classification and TSPAN-relevant analyses. Immunohistochemistry (IHC) and immunofluorescence (IF) stains were conducted to observe the protein expression level. To evaluate the relative biological pathways, gene-set enrichment analysis was performed.

RESULTS

Of the highly expressed TSPAN genes in CRC tissues (TSPAN8, TSPAN29, and TSPAN30), TSPAN8 was notably overexpressed in CMS3-classified primary tissues. The overexpression of TSPAN8 protein in CMS3 CRC was also observed by IHC and IF staining. As a result of gene-set enrichment analysis, TSPAN8 may potentially play a role in organizing signaling complexes for kinase-based metabolic deregulation in CMS3 CRC.

CONCLUSIONS

The present study reports the overexpression of TSPAN8 in CMS3 CRC. This study proposes TSPAN8 as a subtype-specific biomarker for CMS3 CRC. This finding provides a foundation for future CMS-based studies of CRC, a complex disease and the second leading cause of cancer mortality worldwide.

Transcriptomic Profiling of Plasma Extracellular Vesicles Enables Reliable Annotation of the Cancer-Specific Transcriptome and Molecular Subtype

Longitudinal monitoring of patients with advanced cancers is crucial to evaluate both disease burden and treatment response. Current liquid biopsy approaches mostly rely on the detection of DNA-based biomarkers. However, plasma RNA analysis can unleash tremendous opportunities for tumor state interrogation and molecular subtyping. Through the application of deep learning algorithms to the deconvolved transcriptomes of RNA within plasma extracellular vesicles (evRNA), we successfully predicted consensus molecular subtypes in patients with metastatic colorectal cancer. Analysis of plasma evRNA also enabled monitoring of changes in transcriptomic subtype under treatment selection pressure and identification of molecular pathways associated with recurrence. This approach also revealed expressed gene fusions and neoepitopes from evRNA. These results demonstrate the feasibility of using transcriptomic-based liquid biopsy platforms for precision oncology approaches, spanning from the longitudinal monitoring of tumor subtype changes to the identification of expressed fusions and neoantigens as cancer-specific therapeutic targets, sans the need for tissue-based sampling.

SIGNIFICANCE

The development of an approach to interrogate molecular subtypes, cancer-associated pathways, and differentially expressed genes through RNA sequencing of plasma extracellular vesicles lays the foundation for liquid biopsy-based longitudinal monitoring of patient tumor transcriptomes.

Few-shot genes selection: subset of PAM50 genes for breast cancer subtypes classification

BACKGROUND

In recent years, researchers have made significant strides in understanding the heterogeneity of breast cancer and its various subtypes. However, the wealth of genomic and proteomic data available today necessitates efficient frameworks, instruments, and computational tools for meaningful analysis. Despite its success as a prognostic tool, the PAM50 gene signature's reliance on many genes presents challenges in terms of cost and complexity. Consequently, there is a need for more efficient methods to classify breast cancer subtypes using a reduced gene set accurately.

RESULTS

This study explores the potential of achieving precise breast cancer subtype categorization using a reduced gene set derived from the PAM50 gene signature. By employing a "Few-Shot Genes Selection" method, we randomly select smaller subsets from PAM50 and evaluate their performance using metrics and a linear model, specifically the Support Vector Machine (SVM) classifier. In addition, we aim to assess whether a more compact gene set can maintain performance while simplifying the classification process. Our findings demonstrate that certain reduced gene subsets can perform comparable or superior to the full PAM50 gene signature.

CONCLUSIONS

The identified gene subsets, with 36 genes, have the potential to contribute to the development of more cost-effective and streamlined diagnostic tools in breast cancer research and clinical settings.

Integrative analysis of RNA expression data unveils distinct cancer types through machine learning techniques

Cancer is a highly complex and heterogeneous disease. Traditional methods of cancer classification based on histopathology have limitations in guiding personalized prognosis and therapy. Gene expression profiling provides a powerful approach to unraveling molecular intricacies and better-stratifying cancer subtypes. In this study, we performed an integrative analysis of RNA sequencing data from five cancer types - BRCA, KIRC, COAD, LUAD, and PRAD. A machine learning workflow consisting of dataset identification, normalization, feature selection, dimensionality reduction, clustering, and classification was implemented. The k-means algorithm was applied to categorize samples into distinct clusters based solely on gene expression patterns. Five unique clusters emerged from the unsupervised machine learning based analysis, significantly correlating with the known cancer types. BRCA aligned predominantly with one cluster, while COAD spanned three clusters. KIRC was represented within two main clusters. LUAD is associated strongly with a single cluster and PRAD with another cluster. This demonstrates the ability of machine learning approaches to unravel complex signatures within transcriptomic profiles that can delineate cancer subtypes. The proposed study highlights the potential of integrative analytics to derive meaningful biological insights from high-dimensional omics datasets. Molecular subtyping through machine learning clustering enhances our understanding of the intrinsic heterogeneities and pathways dysregulated in different cancers. Overall, this study exemplifies a powerful computational framework to classify gene expressions of patients having different types of cancers and guide personalized therapeutic decisions. Finally, Wide Neural Network demonstrates a significantly higher accuracy, achieving 99.834% on the validation set and an even more impressive 99.995% on the test set.

Trends in orphan medicinal products approvals in the European Union between 2010-2022

BACKGROUND

Over the last twenty years of orphan drug regulation in Europe, the regulatory framework has increased its complexity, with different regulatory paths and tools engineered to facilitate the innovation and accelerate approvals. Recently, the proposal of the new Pharmaceutical Legislation for the European Union, which will replace at least three Regulations and one Directive, was released and its new framework is raising many questions. The aim of this study was to present a characterisation of the Orphan Medicinal Products (OMPs) authorised by the European Commission (EC), between 2010 and 2022, looking into eighteen variables, contributing to the ongoing discussion on the proposal and implementation of the new Pharmaceutical Legislation proposed.

METHODS

Data of the OMPs identified and approved between 2010 and 2022 were extracted from the European Public Assessment Reports (EPARs) produced by the European Medicines Agency. Information regarding legal basis of the application, applicant, protocol assistance received, type of authorization, registration status, type of molecule, ATC code, therapeutic area, target age, disease prevalence, number of pivotal clinical trials supporting the application, clinical trial designs, respective efficacy endpoints and number of patients enrolled in the pivotal clinical trials were extracted. A descriptive statistical analysis was applied.

RESULTS

We identified 192 OMPs approved in the period between 2010 and 2022. 89% of the OMPs have legal basis of "full application". 86% of the sponsors received protocol assistance whereas 64% of the MAA benefited from the accelerated assessment. 53% of the active substances are small molecules; about 1 in 5 molecules are repurposed. 40% of the OMPs have oncological therapeutic indications and 56% of the OMPs are intended to treat only adults. 71% of the products were approved based on a single pivotal trial.

CONCLUSIONS

This analysis of OMPs approved between 2010 and 2022 shows that a shift has occurred in the rare disease medicine development space. Through the period studied we observe an increase of non-small molecules approved, accelerated assessment received and non-standard MA's granted.

[Identification of breast cancer subtypes based on graph convolutional network]

Identification of molecular subtypes of malignant tumors plays a vital role in individualized diagnosis, personalized treatment, and prognosis prediction of cancer patients. The continuous improvement of comprehensive tumor genomics database and the ongoing breakthroughs in deep learning technology have driven further advancements in computer-aided tumor classification. Although the existing classification methods based on gene expression omnibus database take the complexity of cancer molecular classification into account, they ignore the internal correlation and synergism of genes. To solve this problem, we propose a multi-layer graph convolutional network model for breast cancer subtype classification combined with hierarchical attention network. This model constructs the graph embedding datasets of patients' genes, and develops a new end-to-end multi-classification model, which can effectively recognize molecular subtypes of breast cancer. A large number of test data prove the good performance of this new model in the classification of breast cancer subtypes. Compared to the original graph convolutional neural networks and two mainstream graph neural network classification algorithms, the new model has remarkable advantages. The accuracy, weight-F1-score, weight-recall, and weight-precision of our model in seven-category classification has reached 0.851 7, 0.823 5, 0.851 7 and 0.793 6 respectively. In the four-category classification, the results are 0.928 5, 0.894 9, 0.928 5 and 0.865 0 respectively. In addition, compared with the latest breast cancer subtype classification algorithms, the method proposed in this paper also achieved the highest classification accuracy. In summary, the model proposed in this paper may serve as an auxiliary diagnostic technology, providing a reliable option for precise classification of breast cancer subtypes in the future and laying the theoretical foundation for computer-aided tumor classification.

Decoding leukemia at the single-cell level: clonal architecture, classification, microenvironment, and drug resistance

Leukemias are refractory hematological malignancies, characterized by marked intrinsic heterogeneity which poses significant obstacles to effective treatment. However, traditional bulk sequencing techniques have not been able to effectively unravel the heterogeneity among individual tumor cells. With the emergence of single-cell sequencing technology, it has bestowed upon us an unprecedented resolution to comprehend the mechanisms underlying leukemogenesis and drug resistance across various levels, including the genome, epigenome, transcriptome and proteome. Here, we provide an overview of the currently prevalent single-cell sequencing technologies and a detailed summary of single-cell studies conducted on leukemia, with a specific focus on four key aspects: (1) leukemia's clonal architecture, (2) frameworks to determine leukemia subtypes, (3) tumor microenvironment (TME) and (4) the drug-resistant mechanisms of leukemia. This review provides a comprehensive summary of current single-cell studies on leukemia and highlights the markers and mechanisms that show promising clinical implications for the diagnosis and treatment of leukemia.

Exploring extracellular matrix and prostaglandin pathway alterations across varying resection margin distances of right-sided colonic adenocarcinoma

BACKGROUND

Surgical resection followed by indicated adjuvant therapy offers potential curative treatment in colonic adenocarcinoma. Beyond the well-established seed and soil theory of colon cancer progression, the 'normal-appearing' tissues near the tumor are not genuinely normal and remain as remnants in patients following surgery. Our objective was to elucidate the alteration of gene expression and pathways across various distances of resection margins in right-sided colonic adenocarcinoma.

METHODS

Twenty-seven fresh samples of primary cancer and 56 matched non-tumor tissues adjacent to the tumor (NAT) were collected from patients with resectable right-sided colon cancer. NAT were systematically obtained at varying distances (1, 5, and 10 cm) on both proximal and distal sides. Comprehensive gene expression analysis was performed using 770-gene PanCancer Progression Panel, delineating distinctive pathways and functional predictions for each region.

RESULTS

Distinctive gene signatures and pathways exhibited by normal-appearing tissues were discovered at varying distances from cancer. Notably, SFRP2, PTGDS, COL1A1, IL1B, THBS2, PTGIS, COL1A2, NPR1, and BGN were upregulated, while ENPEP, MMP1, and NRCAM were downregulated significantly in 1-cm tissue compared to farther distances. Substantial alterations in the extracellular matrix (ECM) and prostaglandin/thromboxane synthesis were significantly evident at the 1-cm distance. Functional analysis indicated enhanced cell viability and survival, alongside reduced cellular death and apoptosis.

CONCLUSIONS

Different distances exerted a significant impact on gene alteration within the normal-looking mucosa surrounding primary cancer, influenced by various mechanisms. These findings may highlight potential therapeutic targets related to the ECM and prostaglandin/thromboxane pathways for treatment strategies.

Systems Biology in Cancer Diagnosis Integrating Omics Technologies and Artificial Intelligence to Support Physician Decision Making

Cancer is the second major cause of disease-related death worldwide, and its accurate early diagnosis and therapeutic intervention are fundamental for saving the patient's life. Cancer, as a complex and heterogeneous disorder, results from the disruption and alteration of a wide variety of biological entities, including genes, proteins, mRNAs, miRNAs, and metabolites, that eventually emerge as clinical symptoms. Traditionally, diagnosis is based on clinical examination, blood tests for biomarkers, the histopathology of a biopsy, and imaging (MRI, CT, PET, and US). Additionally, omics biotechnologies help to further characterize the genome, metabolome, microbiome traits of the patient that could have an impact on the prognosis and patient's response to the therapy. The integration of all these data relies on gathering of several experts and may require considerable time, and, unfortunately, it is not without the risk of error in the interpretation and therefore in the decision. Systems biology algorithms exploit Artificial Intelligence (AI) combined with omics technologies to perform a rapid and accurate analysis and integration of patient's big data, and support the physician in making diagnosis and tailoring the most appropriate therapeutic intervention. However, AI is not free from possible diagnostic and prognostic errors in the interpretation of images or biochemical-clinical data. Here, we first describe the methods used by systems biology for combining AI with omics and then discuss the potential, challenges, limitations, and critical issues in using AI in cancer research.

Attention-based GCN integrates multi-omics data for breast cancer subtype classification and patient-specific gene marker identification

Breast cancer is a heterogeneous disease and can be divided into several subtypes with unique prognostic and molecular characteristics. The classification of breast cancer subtypes plays an important role in the precision treatment and prognosis of breast cancer. Benefitting from the relation-aware ability of a graph convolution network (GCN), we present a multi-omics integrative method, the attention-based GCN (AGCN), for breast cancer molecular subtype classification using messenger RNA expression, copy number variation and deoxyribonucleic acid methylation multi-omics data. In the extensive comparative studies, our AGCN models outperform state-of-the-art methods under different experimental conditions and both attention mechanisms and the graph convolution subnetwork play an important role in accurate cancer subtype classification. The layer-wise relevance propagation (LRP) algorithm is used for the interpretation of model decision, which can identify patient-specific important biomarkers that are reported to be related to the occurrence and development of breast cancer. Our results highlighted the effectiveness of the GCN and attention mechanisms in multi-omics integrative analysis and the implement of the LRP algorithm can provide biologically reasonable insights into model decision.

Automated exploitation of deep learning for cancer patient stratification across multiple types

MOTIVATION

Recent frameworks based on deep learning have been developed to identify cancer subtypes from high-throughput gene expression profiles. Unfortunately, the performance of deep learning is highly dependent on its neural network architectures which are often hand-crafted with expertise in deep neural networks, meanwhile, the optimization and adjustment of the network are usually costly and time consuming.

RESULTS

To address such limitations, we proposed a fully automated deep neural architecture search model for diagnosing consensus molecular subtypes from gene expression data (DNAS). The proposed model uses ant colony algorithm, one of the heuristic swarm intelligence algorithms, to search and optimize neural network architecture, and it can automatically find the optimal deep learning model architecture for cancer diagnosis in its search space. We validated DNAS on eight colorectal cancer datasets, achieving the average accuracy of 95.48%, the average specificity of 98.07%, and the average sensitivity of 96.24%, respectively. Without the loss of generality, we investigated the general applicability of DNAS further on other cancer types from different platforms including lung cancer and breast cancer, and DNAS achieved an area under the curve of 95% and 96%, respectively. In addition, we conducted gene ontology enrichment and pathological analysis to reveal interesting insights into cancer subtype identification and characterization across multiple cancer types.

AVAILABILITY AND IMPLEMENTATION

The source code and data can be downloaded from https://github.com/userd113/DNAS-main. And the web server of DNAS is publicly accessible at 119.45.145.120:5001.

Deep learning on electronic medical records identifies distinct subphenotypes of diabetic kidney disease driven by genetic variations in the Rho pathway

Kidney disease affects 50% of all diabetic patients; however, prediction of disease progression has been challenging due to inherent disease heterogeneity. We use deep learning to identify novel genetic signatures prognostically associated with outcomes. Using autoencoders and unsupervised clustering of electronic health record data on 1,372 diabetic kidney disease patients, we establish two clusters with differential prevalence of end-stage kidney disease. Exome-wide associations identify a novel variant in ARHGEF18, a Rho guanine exchange factor specifically expressed in glomeruli. Overexpression of ARHGEF18 in human podocytes leads to impairments in focal adhesion architecture, cytoskeletal dynamics, cellular motility, and RhoA/Rac1 activation. Mutant GEF18 is resistant to ubiquitin mediated degradation leading to pathologically increased protein levels. Our findings uncover the first known disease-causing genetic variant that affects protein stability of a cytoskeletal regulator through impaired degradation, a potentially novel class of expression quantitative trait loci that can be therapeutically targeted.

Senescence-based colorectal cancer subtyping reveals distinct molecular characteristics and therapeutic strategies

Cellular senescence has been listed as a hallmark of cancer, but its role in colorectal cancer (CRC) remains unclear. We comprehensively evaluated the transcriptome, genome, digital pathology, and clinical data from multiple datasets of CRC patients and proposed a novel senescence subtype for CRC. Multi-omics data was used to analyze the biological features, tumor microenvironment, and mutation landscape of senescence subtypes, as well as drug sensitivity and immunotherapy response. The senescence score was constructed to better quantify senescence in each patient for clinical use. Unsupervised learning revealed three transcriptome-based senescence subtypes. Cluster 1, characterized by low senescence and activated proliferative pathways, was sensitive to chemotherapeutic drugs. Cluster 2, characterized by intermediate senescence and high immune infiltration, exhibited significant immunotherapeutic advantages. Cluster 3, characterized by high senescence, high immune, and stroma infiltration, had a worse prognosis and maybe benefit from targeted therapy. We further constructed a senescence scoring system based on seven senescent genes through machine learning. Lower senescence scores were highly predictive of longer disease-free survival, and patients with low senescence scores may benefit from immunotherapy. We proposed the senescence subtypes of CRC and our findings provide potential treatment interventions for each CRC senescence subtype to promote precision treatment.

Artificial intelligence in colorectal surgery: an AI-powered systematic review

Artificial intelligence (AI) has the potential to revolutionize surgery in the coming years. Still, it is essential to clarify what the meaningful current applications are and what can be reasonably expected. This AI-powered review assessed the role of AI in colorectal surgery. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)-compliant systematic search of PubMed, Embase, Scopus, Cochrane Library databases, and gray literature was conducted on all available articles on AI in colorectal surgery (from January 1 1997 to March 1 2021), aiming to define the perioperative applications of AI. Potentially eligible studies were identified using novel software powered by natural language processing (NLP) and machine learning (ML) technologies dedicated to systematic reviews. Out of 1238 articles identified, 115 were included in the final analysis. Available articles addressed the role of AI in several areas of interest. In the preoperative phase, AI can be used to define tailored treatment algorithms, support clinical decision-making, assess the risk of complications, and predict surgical outcomes and survival. Intraoperatively, AI-enhanced surgery and integration of AI in robotic platforms have been suggested. After surgery, AI can be implemented in the Enhanced Recovery after Surgery (ERAS) pathway. Additional areas of applications included the assessment of patient-reported outcomes, automated pathology assessment, and research. Available data on these aspects are limited, and AI in colorectal surgery is still in its infancy. However, the rapid evolution of technologies makes it likely that it will increasingly be incorporated into everyday practice.

The role of deep learning in diagnosing colorectal cancer

Colon cancer is a major public health issue, affecting a growing number of individuals worldwide. Proper and early diagnosis of colon cancer is the necessary first step toward effective treatment and/or prevention of future disease relapse. Artificial intelligence and its subtypes, deep learning in particular, tend nowadays to have an expanding role in all fields of medicine, and diagnosing colon cancer is no exception. This report aims to summarize the entire application spectrum of deep learning in all diagnostic tests regarding colon cancer, from endoscopy and histologic examination to medical imaging and screening serologic tests.

Comparison of cancer subtype identification methods combined with feature selection methods in omics data analysis

BACKGROUND

Cancer subtype identification is important for the early diagnosis of cancer and the provision of adequate treatment. Prior to identifying the subtype of cancer in a patient, feature selection is also crucial for reducing the dimensionality of the data by detecting genes that contain important information about the cancer subtype. Numerous cancer subtyping methods have been developed, and their performance has been compared. However, combinations of feature selection and subtype identification methods have rarely been considered. This study aimed to identify the best combination of variable selection and subtype identification methods in single omics data analysis.

RESULTS

Combinations of six filter-based methods and six unsupervised subtype identification methods were investigated using The Cancer Genome Atlas (TCGA) datasets for four cancers. The number of features selected varied, and several evaluation metrics were used. Although no single combination was found to have a distinctively good performance, Consensus Clustering (CC) and Neighborhood-Based Multi-omics Clustering (NEMO) used with variance-based feature selection had a tendency to show lower p-values, and nonnegative matrix factorization (NMF) stably showed good performance in many cases unless the Dip test was used for feature selection. In terms of accuracy, the combination of NMF and similarity network fusion (SNF) with Monte Carlo Feature Selection (MCFS) and Minimum-Redundancy Maximum Relevance (mRMR) showed good overall performance. NMF always showed among the worst performances without feature selection in all datasets, but performed much better when used with various feature selection methods. iClusterBayes (ICB) had decent performance when used without feature selection.

CONCLUSIONS

Rather than a single method clearly emerging as optimal, the best methodology was different depending on the data used, the number of features selected, and the evaluation method. A guideline for choosing the best combination method under various situations is provided.

Mutation-Attention (MuAt): deep representation learning of somatic mutations for tumour typing and subtyping

BACKGROUND

Cancer genome sequencing enables accurate classification of tumours and tumour subtypes. However, prediction performance is still limited using exome-only sequencing and for tumour types with low somatic mutation burden such as many paediatric tumours. Moreover, the ability to leverage deep representation learning in discovery of tumour entities remains unknown.

METHODS

We introduce here Mutation-Attention (MuAt), a deep neural network to learn representations of simple and complex somatic alterations for prediction of tumour types and subtypes. In contrast to many previous methods, MuAt utilizes the attention mechanism on individual mutations instead of aggregated mutation counts.

RESULTS

We trained MuAt models on 2587 whole cancer genomes (24 tumour types) from the Pan-Cancer Analysis of Whole Genomes (PCAWG) and 7352 cancer exomes (20 types) from the Cancer Genome Atlas (TCGA). MuAt achieved prediction accuracy of 89% for whole genomes and 64% for whole exomes, and a top-5 accuracy of 97% and 90%, respectively. MuAt models were found to be well-calibrated and perform well in three independent whole cancer genome cohorts with 10,361 tumours in total. We show MuAt to be able to learn clinically and biologically relevant tumour entities including acral melanoma, SHH-activated medulloblastoma, SPOP-associated prostate cancer, microsatellite instability, POLE proofreading deficiency, and MUTYH-associated pancreatic endocrine tumours without these tumour subtypes and subgroups being provided as training labels. Finally, scrunity of MuAt attention matrices revealed both ubiquitous and tumour-type specific patterns of simple and complex somatic mutations.

CONCLUSIONS

Integrated representations of somatic alterations learnt by MuAt were able to accurately identify histological tumour types and identify tumour entities, with potential to impact precision cancer medicine.

AttOmics: attention-based architecture for diagnosis and prognosis from omics data

MOTIVATION

The increasing availability of high-throughput omics data allows for considering a new medicine centered on individual patients. Precision medicine relies on exploiting these high-throughput data with machine-learning models, especially the ones based on deep-learning approaches, to improve diagnosis. Due to the high-dimensional small-sample nature of omics data, current deep-learning models end up with many parameters and have to be fitted with a limited training set. Furthermore, interactions between molecular entities inside an omics profile are not patient specific but are the same for all patients.

RESULTS

In this article, we propose AttOmics, a new deep-learning architecture based on the self-attention mechanism. First, we decompose each omics profile into a set of groups, where each group contains related features. Then, by applying the self-attention mechanism to the set of groups, we can capture the different interactions specific to a patient. The results of different experiments carried out in this article show that our model can accurately predict the phenotype of a patient with fewer parameters than deep neural networks. Visualizing the attention maps can provide new insights into the essential groups for a particular phenotype.

AVAILABILITY AND IMPLEMENTATION

The code and data are available at https://forge.ibisc.univ-evry.fr/abeaude/AttOmics. TCGA data can be downloaded from the Genomic Data Commons Data Portal.

The NF-Y splicing signature controls hybrid EMT and ECM-related pathways to promote aggressiveness of colon cancer

Aberrant splicing events are associated with colorectal cancer (CRC) and provide new opportunities for tumor diagnosis and treatment. The expression of the splice variants of NF-YA, the DNA binding subunit of the transcription factor NF-Y, is deregulated in multiple cancer types compared to healthy tissues. NF-YAs and NF-YAl isoforms differ in the transactivation domain, which may result in distinct transcriptional programs. In this study, we demonstrated that the NF-YAl transcript is higher in aggressive mesenchymal CRCs and predicts shorter patients' survival. In 2D and 3D conditions, CRC cells overexpressing NF-YAl (NF-YAl^high) exhibit reduced cell proliferation, rapid single cell amoeboid-like migration, and form irregular spheroids with poor cell-to-cell adhesion. Compared to NF-YAs^high, NF-YAl^high cells show changes in the transcription of genes involved in epithelial-mesenchymal transition, extracellular matrix and cell adhesion. NF-YAl and NF-YAs bind similarly to the promoter of the E-cadherin gene, but oppositely regulate its transcription. The increased metastatic potential of NF-YAl^high cells in vivo was confirmed in zebrafish xenografts. These results suggest that the NF-YAl splice variant could be a new CRC prognostic factor and that splice-switching strategies may reduce metastatic CRC progression.

Learning vector quantized representation for cancer subtypes identification

BACKGROUND AND OBJECTIVE

Defining and separating cancer subtypes is essential for facilitating personalized therapy modality and prognosis of patients. The definition of subtypes has been constantly recalibrated as a result of our deepened understanding. During this recalibration, researchers often rely on clustering of cancer data to provide an intuitive visual reference that could reveal the intrinsic characteristics of subtypes. The data being clustered are often omics data such as transcriptomics that have strong correlations to the underlying biological mechanism. However, while existing studies have shown promising results, they suffer from issues associated with omics data: sample scarcity and high dimensionality while they impose unrealistic assumptions to extract useful features from the data while avoiding overfitting to spurious correlations.

METHODS

This paper proposes to leverage a recent strong generative model, Vector-Quantized Variational AutoEncoder, to tackle the data issues and extract discrete representations that are crucial to the quality of subsequent clustering by retaining only information relevant to reconstructing the input.

RESULTS

Extensive experiments and medical analysis on multiple datasets comprising 10 distinct cancers demonstrate the proposed clustering results can significantly and robustly improve prognosis over prevalent subtyping systems.

CONCLUSION

Our proposal does not impose strict assumptions on data distribution; while, its latent features are better representations of the transcriptomic data in different cancer subtypes, capable of yielding superior clustering performance with any mainstream clustering method.

SUPREME: multiomics data integration using graph convolutional networks

To pave the road towards precision medicine in cancer, patients with similar biology ought to be grouped into same cancer subtypes. Utilizing high-dimensional multiomics datasets, integrative approaches have been developed to uncover cancer subtypes. Recently, Graph Neural Networks have been discovered to learn node embeddings utilizing node features and associations on graph-structured data. Some integrative prediction tools have been developed leveraging these advances on multiple networks with some limitations. Addressing these limitations, we developed SUPREME, a node classification framework, which integrates multiple data modalities on graph-structured data. On breast cancer subtyping, unlike existing tools, SUPREME generates patient embeddings from multiple similarity networks utilizing multiomics features and integrates them with raw features to capture complementary signals. On breast cancer subtype prediction tasks from three datasets, SUPREME outperformed other tools. SUPREME-inferred subtypes had significant survival differences, mostly having more significance than ground truth, and outperformed nine other approaches. These results suggest that with proper multiomics data utilization, SUPREME could demystify undiscovered characteristics in cancer subtypes that cause significant survival differences and could improve ground truth label, which depends mainly on one datatype. In addition, to show model-agnostic property of SUPREME, we applied it to two additional datasets and had a clear outperformance.

MODILM: towards better complex diseases classification using a novel multi-omics data integration learning model

BACKGROUND

Accurately classifying complex diseases is crucial for diagnosis and personalized treatment. Integrating multi-omics data has been demonstrated to enhance the accuracy of analyzing and classifying complex diseases. This can be attributed to the highly correlated nature of the data with various diseases, as well as the comprehensive and complementary information it provides. However, integrating multi-omics data for complex diseases is challenged by data characteristics such as high imbalance, scale variation, heterogeneity, and noise interference. These challenges further emphasize the importance of developing effective methods for multi-omics data integration.

RESULTS

We proposed a novel multi-omics data learning model called MODILM, which integrates multiple omics data to improve the classification accuracy of complex diseases by obtaining more significant and complementary information from different single-omics data. Our approach includes four key steps: 1) constructing a similarity network for each omics data using the cosine similarity measure, 2) leveraging Graph Attention Networks to learn sample-specific and intra-association features from similarity networks for single-omics data, 3) using Multilayer Perceptron networks to map learned features to a new feature space, thereby strengthening and extracting high-level omics-specific features, and 4) fusing these high-level features using a View Correlation Discovery Network to learn cross-omics features in the label space, which results in unique class-level distinctiveness for complex diseases. To demonstrate the effectiveness of MODILM, we conducted experiments on six benchmark datasets consisting of miRNA expression, mRNA, and DNA methylation data. Our results show that MODILM outperforms state-of-the-art methods, effectively improving the accuracy of complex disease classification.

CONCLUSIONS

Our MODILM provides a more competitive way to extract and integrate important and complementary information from multiple omics data, providing a very promising tool for supporting decision-making for clinical diagnosis.

moBRCA-net: a breast cancer subtype classification framework based on multi-omics attention neural networks

BACKGROUND

Breast cancer is a highly heterogeneous disease that comprises multiple biological components. Owing its diversity, patients have different prognostic outcomes; hence, early diagnosis and accurate subtype prediction are critical for treatment. Standardized breast cancer subtyping systems, mainly based on single-omics datasets, have been developed to ensure proper treatment in a systematic manner. Recently, multi-omics data integration has attracted attention to provide a comprehensive view of patients but poses a challenge due to the high dimensionality. In recent years, deep learning-based approaches have been proposed, but they still present several limitations.

RESULTS

In this study, we describe moBRCA-net, an interpretable deep learning-based breast cancer subtype classification framework that uses multi-omics datasets. Three omics datasets comprising gene expression, DNA methylation and microRNA expression data were integrated while considering the biological relationships among them, and a self-attention module was applied to each omics dataset to capture the relative importance of each feature. The features were then transformed to new representations considering the respective learned importance, allowing moBRCA-net to predict the subtype.

CONCLUSIONS

Experimental results confirmed that moBRCA-net has a significantly enhanced performance compared with other methods, and the effectiveness of multi-omics integration and omics-level attention were identified. moBRCA-net is publicly available at https://github.com/cbi-bioinfo/moBRCA-net .

meth-SemiCancer: a cancer subtype classification framework via semi-supervised learning utilizing DNA methylation profiles

BACKGROUND

Identification of the cancer subtype plays a crucial role to provide an accurate diagnosis and proper treatment to improve the clinical outcomes of patients. Recent studies have shown that DNA methylation is one of the key factors for tumorigenesis and tumor growth, where the DNA methylation signatures have the potential to be utilized as cancer subtype-specific markers. However, due to the high dimensionality and the low number of DNA methylome cancer samples with the subtype information, still, to date, a cancer subtype classification method utilizing DNA methylome datasets has not been proposed.

RESULTS

In this paper, we present meth-SemiCancer, a semi-supervised cancer subtype classification framework based on DNA methylation profiles. The proposed model was first pre-trained based on the methylation datasets with the cancer subtype labels. After that, meth-SemiCancer generated the pseudo-subtypes for the cancer datasets without subtype information based on the model's prediction. Finally, fine-tuning was performed utilizing both the labeled and unlabeled datasets.

CONCLUSIONS

From the performance comparison with the standard machine learning-based classifiers, meth-SemiCancer achieved the highest average F1-score and Matthews correlation coefficient, outperforming other methods. Fine-tuning the model with the unlabeled patient samples by providing the proper pseudo-subtypes, encouraged meth-SemiCancer to generalize better than the supervised neural network-based subtype classification method. meth-SemiCancer is publicly available at https://github.com/cbi-bioinfo/meth-SemiCancer .

Improved Delineation of Colorectal Cancer Molecular Subtypes and Functional Profiles with a 62-Gene Panel

Since its establishment in 2015, the transcriptomics-based consensus molecular subtype (CMS) classification has unified our understanding of colorectal cancer. Each of the four CMS exhibited distinctive high-level molecular signatures that correlated well with prognosis and treatment response. Nonetheless, many key aspects of colorectal cancer progression and intra-subtype heterogeneity remain unresolved. This is partly because the bulk transcriptomic data used to define CMS contain substantial interference from non-tumor cells. Here, we propose a concise panel of 62 genes that not only accurately recapitulates all key characteristics of the four original CMS but also identifies three additional subpopulations with unique molecular signatures. Validation on independent cohorts confirms that the new CMS4 intra-subtypes coincide with single-cell-derived intrinsic subtypes and that the panel consists of many immune cell-type markers that can capture the status of tumor microenvironment. Furthermore, a 2D embedding of CMS structure based on the proposed gene panel provides a high-resolution view of the functional pathways and cell-type markers that underlie each CMS intra-subtype and the continuous progression from CMS2 to CMS4 subtypes. Our gene panel and 2D visualization refined the delineation of colorectal cancer subtypes and could aid further discovery of molecular mechanisms in colorectal cancer.

IMPLICATIONS

: Well-selected gene panel and representation can capture both the continuum of cancer cell states and tumor microenvironment status.

MmCMS: mouse models' consensus molecular subtypes of colorectal cancer

BACKGROUND

Colorectal cancer (CRC) primary tumours are molecularly classified into four consensus molecular subtypes (CMS1-4). Genetically engineered mouse models aim to faithfully mimic the complexity of human cancers and, when appropriately aligned, represent ideal pre-clinical systems to test new drug treatments. Despite its importance, dual-species classification has been limited by the lack of a reliable approach. Here we utilise, develop and test a set of options for human-to-mouse CMS classifications of CRC tissue.

METHODS

Using transcriptional data from established collections of CRC tumours, including human (TCGA cohort; n = 577) and mouse (n = 57 across n = 8 genotypes) tumours with combinations of random forest and nearest template prediction algorithms, alongside gene ontology collections, we comprehensively assess the performance of a suite of new dual-species classifiers.

RESULTS

We developed three approaches: MmCMS-A; a gene-level classifier, MmCMS-B; an ontology-level approach and MmCMS-C; a combined pathway system encompassing multiple biological and histological signalling cascades. Although all options could identify tumours associated with stromal-rich CMS4-like biology, MmCMS-A was unable to accurately classify the biology underpinning epithelial-like subtypes (CMS2/3) in mouse tumours.

CONCLUSIONS

When applying human-based transcriptional classifiers to mouse tumour data, a pathway-level classifier, rather than an individual gene-level system, is optimal. Our R package enables researchers to select suitable mouse models of human CRC subtype for their experimental testing.

Using Deep Learning with Bayesian–Gaussian Inspired Convolutional Neural Architectural Search for Cancer Recognition and Classification from Histopathological Image Frames

We propose a neural architectural search model which examines histopathological images to detect the presence of cancer in both lung and colon tissues. In recent times, deep artificial neural networks have made tremendous impacts in healthcare. However, obtaining an optimal artificial neural network model that could yield excellent performance during training, evaluation, and inferencing has been a bottleneck for researchers. Our method uses a Bayesian convolutional neural architectural search algorithm in collaboration with Gaussian processes to provide an efficient neural network architecture for efficient colon and lung cancer classification and recognition. The proposed model learns by using the Gaussian process to estimate the required optimal architectural values by choosing a set of model parameters through the exploitation of the expected improvement (EI) values, thereby minimizing the number of sampled trials and suggesting the best model architecture. Several experiments were conducted, and a landmark performance was obtained in both validation and test data through the evaluation of the proposed model on a dataset consisting of 25,000 images of five different classes with convergence and F1-score matrices.

The KRAS-Mutant Consensus Molecular Subtype 3 Reveals an Immunosuppressive Tumor Microenvironment in Colorectal Cancer

Colorectal cancers (CRC) with KRAS mutations (KRAS^mut) are frequently included in consensus molecular subtype 3 (CMS3) with profound metabolic deregulation. We explored the transcriptomic impact of KRAS^mut, focusing on the tumor microenvironment (TME) and pathways beyond metabolic deregulation. The status of KRAS^mut in patients with CRC was investigated and overall survival (OS) was compared with wild-type KRAS (KRAS^wt). Next, we identified CMS, and further investigated differentially expressed genes (DEG) of KRAS^mut and distinctive pathways. Lastly, we used spatially resolved gene expression profiling to define the effect of KRAS^mut in the TME regions of CMS3-classified CRC tissues. CRC patients with KRAS^mut were mainly enriched in CMS3. Their specific enrichments of immune gene signatures in immunosuppressive TME were associated with worse OS. Activation of TGFβ signaling by KRAS^mut was related to reduced pro-inflammatory and cytokine gene signatures, leading to suppression of immune infiltration. Digital spatial profiling in TME regions of KRAS^mut CMS3-classified tissues suggested up-regulated genes, CD40, CTLA4, ARG1, STAT3, IDO, and CD274, that could be characteristic of immune suppression in TME. This study may help to depict the complex transcriptomic profile of KRAS^mut in immunosuppressive TME. Future studies and clinical trials in CRC patients with KRAS^mut should consider these transcriptional landscapes.

AI-DrugNet: A network-based deep learning model for drug repurposing and combination therapy in neurological disorders

Discovering effective therapies is difficult for neurological and developmental disorders in that disease progression is often associated with a complex and interactive mechanism. Over the past few decades, few drugs have been identified for treating Alzheimer's disease (AD), especially for impacting the causes of cell death in AD. Although drug repurposing is gaining more success in developing therapeutic efficacy for complex diseases such as common cancer, the complications behind AD require further study. Here, we developed a novel prediction framework based on deep learning to identify potential repurposed drug therapies for AD, and more importantly, our framework is broadly applicable and may generalize to identifying potential drug combinations in other diseases. Our prediction framework is as follows: we first built a drug-target pair (DTP) network based on multiple drug features and target features, as well as the associations between DTP nodes where drug-target pairs are the DTP nodes and the associations between DTP nodes are represented as the edges in the AD disease network; furthermore, we incorporated the drug-target feature from the DTP network and the relationship information between drug-drug, target-target, drug-target within and outside of drug-target pairs, representing each drug-combination as a quartet to generate corresponding integrated features; finally, we developed an AI-based Drug discovery Network (AI-DrugNet), which exhibits robust predictive performance. The implementation of our network model help identify potential repurposed and combination drug options that may serve to treat AD and other diseases.

Machine Learning Methods for Cancer Classification Using Gene Expression Data: A Review

Cancer is a term that denotes a group of diseases caused by the abnormal growth of cells that can spread in different parts of the body. According to the World Health Organization (WHO), cancer is the second major cause of death after cardiovascular diseases. Gene expression can play a fundamental role in the early detection of cancer, as it is indicative of the biochemical processes in tissue and cells, as well as the genetic characteristics of an organism. Deoxyribonucleic acid (DNA) microarrays and ribonucleic acid (RNA)-sequencing methods for gene expression data allow quantifying the expression levels of genes and produce valuable data for computational analysis. This study reviews recent progress in gene expression analysis for cancer classification using machine learning methods. Both conventional and deep learning-based approaches are reviewed, with an emphasis on the application of deep learning models due to their comparative advantages for identifying gene patterns that are distinctive for various types of cancers. Relevant works that employ the most commonly used deep neural network architectures are covered, including multi-layer perceptrons, as well as convolutional, recurrent, graph, and transformer networks. This survey also presents an overview of the data collection methods for gene expression analysis and lists important datasets that are commonly used for supervised machine learning for this task. Furthermore, we review pertinent techniques for feature engineering and data preprocessing that are typically used to handle the high dimensionality of gene expression data, caused by a large number of genes present in data samples. The paper concludes with a discussion of future research directions for machine learning-based gene expression analysis for cancer classification.

A survey on gene expression data analysis using deep learning methods for cancer diagnosis

Gene Expression Data is the biological data to extract meaningful hidden information from the gene dataset. This gene information is used for disease diagnosis especially in cancer treatment based on the variations in gene expression levels. DNA microarray is an efficient method for gene expression classification and prediction of cancer disease for specific types of cancer. Due to the abundance of computing power, deep learning (DL) has become a widespread technique in the healthcare sector. The gene expression dataset has a limited number of samples but a large number of features. Data augmentation is needed for gene expression datasets to overcome the dimensionality problem in gene data. It is a technique to generating the synthetic samples to increase the diversity of data. Deep learning methods are designed to learn and extract the features that come from the raw input data in the form of multidimensional arrays. This paper reviews the existing research in deep learning techniques like Feed Forward Neural Network (FFN), Convolutional Neural Network (CNN), Autoencoder (AE) and Recurrent Neural Network (RNN) for the classification and prediction of cancer disease and its types through gene expression data analysis.

hTERT Gene Modification Using CRISPR-dCas9-dnmt3a System as a Therapeutic Approach Against Glioma

Background

Abnormal DNA methylation patterns have been reported in various diseases, including different cancers. CRISPR/Cas9 is a low-cost and highly effective gene editing tool that has lately revolutionized biotechnology. Studies have shown that the CRISPR/Cas9 system can effectively target and correct methylation.

Objectives

Telomerase plays a survival role for cancer cells. It is encoded by the hTERT gene. The effectiveness of CRISPR/Cas9 in targeting hTERT to treat glioma cancer cells was assessed in this study.

Methods

EF1a-hsaCas9-U6-gRNA vector carrying sgRNA and Cas9 hybrids were used to transfect U87 glioma cells. Four and eight μg/mL polybrene concentrations were investigated to improve transfection efficiency. The expression level of hTERT that has undergone metabisulfite modification was assessed using real-time PCR. Flow cytometry and Western blotting were also used to determine whether telomerase was present in the cells. High-resolution melting analysis (HRM) was used to examine the hTERT promoter's methylation. Finally, flow cytometry was used to measure the apoptotic rate of transfected U87 cells.

Results

The findings demonstrated that gRNA significantly boosted transfection effectiveness. Significant variations were seen in the expression of hTERT in U87 cells at 4 μg/mL polybrene and 80 μg/mL transfection compared to transfection without gRNA and basal cells. Flow cytometry showed a decrease in hTERT levels in transfected cells. Furthermore, transfection with gRNA increased U87 cell apoptosis compared to transfection without gRNA.

Conclusions

It appears that the designed CRISPR/Cas9 system can reduce hTERT expression and telomerase activity and thus inhibit glioma cell growth.

Role of AI and digital pathology for colorectal immuno-oncology

Immunotherapy deals with therapeutic interventions to arrest the progression of tumours using the immune system. These include checkpoint inhibitors, T-cell manipulation, cytokines, oncolytic viruses and tumour vaccines. In this paper, we present a survey of the latest developments on immunotherapy in colorectal cancer (CRC) and the role of artificial intelligence (AI) in this context. Among these, microsatellite instability (MSI) is perhaps the most popular IO biomarker globally. We first discuss the MSI status of tumours, its implications for patient management, and its relationship to immune response. In recent years, several aspiring studies have used AI to predict the MSI status of patients from digital whole-slide images (WSIs) of routine diagnostic slides. We present a survey of AI literature on the prediction of MSI and tumour mutation burden from digitised WSIs of haematoxylin and eosin-stained diagnostic slides. We discuss AI approaches in detail and elaborate their contributions, limitations and key takeaways to drive future research. We further expand this survey to other IO-related biomarkers like immune cell infiltrates and alternate data modalities like immunohistochemistry and gene expression. Finally, we underline possible future directions in immunotherapy for CRC and promise of AI to accelerate this exploration for patient benefits.

A pairwise strategy for imputing predictive features when combining multiple datasets

MOTIVATION

In the training of predictive models using high-dimensional genomic data, multiple studies' worth of data are often combined to increase sample size and improve generalizability. A drawback of this approach is that there may be different sets of features measured in each study due to variations in expression measurement platform or technology. It is often common practice to work only with the intersection of features measured in common across all studies, which results in the blind discarding of potentially useful feature information that is measured in individual or subsets of studies.

RESULTS

We characterize the loss in predictive performance incurred by using only the intersection of feature information available across all studies when training predictors using gene expression data from microarray and sequencing datasets. We study the properties of linear and polynomial regression for imputing discarded features and demonstrate improvements in the external performance of prediction functions through simulation and in gene expression data collected on breast cancer patients. To improve this process, we propose a pairwise strategy that applies any imputation algorithm to two studies at a time and averages imputed features across pairs. We demonstrate that the pairwise strategy is preferable to first merging all datasets together and imputing any resulting missing features. Finally, we provide insights on which subsets of intersected and study-specific features should be used so that missing-feature imputation best promotes cross-study replicability.

AVAILABILITY AND IMPLEMENTATION

The code is available at https://github.com/YujieWuu/Pairwise_imputation.

SUPPLEMENTARY INFORMATION

Supplementary information is available at Bioinformatics online.

DeepCIA: a novel deep-learning model for cancer type identification using class activation map via transcription factor expression

Deep learning methods are powerful analytical tools for large-scale data analysis. Here, we introduce DeepCIA as a novel diagnostic deep-learning model for cancer type identification using a class activation map via transcription factor expression. Although many deep learning researches attempts have recently been made in relation to cancer diagnosis, there are difficulties in using cancer data due to a large-scale problem. Therefore, From The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) public databases, we selected transcription factor expression profiles of eight cancer types. TCGA included 3496 samples and divided the train and validation sets in an 8:2 ratio. ICGC included 552 samples and was used as a test set for external validation. To compare the performance of 1D-CNN models, we also used SVM and KNN from machine learning. In external validation, 1D-CNN showed a high average accuracy of 98% and was superior to support vector machine (SVM) and k-nearest neighbor (KNN) with a difference in the accuracy of 10-12%. Also, 1D-CNN performed very well in several performance metrics (98.2% Recall, 98.1% Precision, 98.2% F score, 99.8% Specificity, 99.8% AUC, and 99.0% Balanced Accuracy). In each data set evaluation, 1-network, 5-network, and 2-network with high accuracy were selected and visualized through the Class Activation Map. We identified the Cys2Hys2 zinc finger group with the highest distribution across all cancer types. Collectively, DeepCIA can be used as a decision support system for cancer and a classifier for diagnosing unknown primary cancer, while emphasizing its usefulness in cancer diagnosis.

CT-based radiogenomic analysis dissects intratumor heterogeneity and predicts prognosis of colorectal cancer: a multi-institutional retrospective study

BACKGROUND

This study aimed to develop a radiogenomic prognostic prediction model for colorectal cancer (CRC) by investigating the biological and clinical relevance of intratumoural heterogeneity.

METHODS

This retrospective multi-cohort study was conducted in three steps. First, we identified genomic subclones using unsupervised deconvolution analysis. Second, we established radiogenomic signatures to link radiomic features with prognostic subclone compositions in an independent radiogenomic dataset containing matched imaging and gene expression data. Finally, the prognostic value of the identified radiogenomic signatures was validated using two testing datasets containing imaging and survival information collected from separate medical centres.

RESULTS

This multi-institutional retrospective study included 1601 patients (714 females and 887 males; mean age, 65 years ± 14 [standard deviation]) with CRC from 5 datasets. Molecular heterogeneity was identified using unsupervised deconvolution analysis of gene expression data. The relative prevalence of the two subclones associated with cell cycle and extracellular matrix pathways identified patients with significantly different survival outcomes. A radiogenomic signature-based predictive model significantly stratified patients into high- and low-risk groups with disparate disease-free survival (HR = 1.74, P = 0.003). Radiogenomic signatures were revealed as an independent predictive factor for CRC by multivariable analysis (HR = 1.59, 95% CI:1.03-2.45, P = 0.034). Functional analysis demonstrated that the 11 radiogenomic signatures were predominantly associated with extracellular matrix and immune-related pathways.

CONCLUSIONS

The identified radiogenomic signatures might be a surrogate for genomic signatures and could complement the current prognostic strategies.

Predicting the influence of Circ_0059706 expression on prognosis in patients with acute myeloid leukemia using classical statistics and machine learning

Background: Various circular RNA (circRNA) molecules are abnormally expressed in acute myeloid leukemia (AML), and associated with disease occurrence and development, as well as patient prognosis. The roles of circ_0059706, a circRNA derived from ID1, in AML remain largely unclear.

Results: Here, we reported circ_0059706 expression in de novo AML and its association with prognosis. We found that circ_0059706 expression was significantly lower in AML patients than in controls (p < 0.001). Survival analysis of patients with AML divided into two groups according to high and low circ_0059706 expression showed that overall survival (OS) of patients with high circ_0059706 expression was significantly longer than that of those with low expression (p < 0.05). Further, female patients with AML and those aged >60 years old in the high circ_0059706 expression group had longer OS than male patients and those younger than 60 years. Multiple regression analysis showed that circ_0059706 was an independent factor-affecting prognosis of all patients with AML. To evaluate the prospects for application of circ_0059706 in machine learning predictions, we developed seven types of algorithm. The gradient boosting (GB) model exhibited higher performance in prediction of 1-year prognosis and 3-year prognosis, with AUROC 0.796 and 0.847. We analyzed the importance of variables and found that circ_0059706 expression level was the first important variables among all 26 factors included in the GB algorithm, suggesting the importance of circ_0059706 in prediction model. Further, overexpression of circ_0059706 inhibited cell growth and increased apoptosis of leukemia cells in vitro.

Conclusion: These results provide evidence that high expression of circ_0059706 is propitious for patient prognosis and suggest circ_0059706 as a potential new biomarker for diagnosis and prognosis evaluation in AML, with high predictive value and good prospects for application in machine learning algorithms.

Deep learning techniques for cancer classification using microarray gene expression data

Cancer is one of the top causes of death globally. Recently, microarray gene expression data has been used to aid in cancer’s effective and early detection. The use of DNA microarray technology to uncover information from the expression levels of thousands of genes has enormous promise. The DNA microarray technique can determine the levels of thousands of genes simultaneously in a single experiment. The analysis of gene expression is critical in many disciplines of biological study to obtain the necessary information. This study analyses all the research studies focused on optimizing gene selection for cancer detection using artificial intelligence. One of the most challenging issues is figuring out how to extract meaningful information from massive databases. Deep Learning architectures have performed efficiently in numerous sectors and are used to diagnose many other chronic diseases and to assist physicians in making medical decisions. In this study, we have evaluated the results of different optimizers on a RNA sequence dataset. The Deep learning algorithm proposed in the study classifies five different forms of cancer, including kidney renal clear cell carcinoma (KIRC), Breast Invasive Carcinoma (BRCA), lung adenocarcinoma (LUAD), Prostate Adenocarcinoma (PRAD) and Colon Adenocarcinoma (COAD). The performance of different optimizers like Stochastic gradient descent (SGD), Root Mean Squared Propagation (RMSProp), Adaptive Gradient Optimizer (AdaGrad), and Adaptive Momentum (AdaM). The experimental results gathered on the dataset affirm that AdaGrad and Adam. Also, the performance analysis has been done using different learning rates and decay rates. This study discusses current advancements in deep learning-based gene expression data analysis using optimized feature selection methods.

omicsGAT: Graph Attention Network for Cancer Subtype Analyses

The use of high-throughput omics technologies is becoming increasingly popular in all facets of biomedical science. The mRNA sequencing (RNA-seq) method reports quantitative measures of more than tens of thousands of biological features. It provides a more comprehensive molecular perspective of studied cancer mechanisms compared to traditional approaches. Graph-based learning models have been proposed to learn important hidden representations from gene expression data and network structure to improve cancer outcome prediction, patient stratification, and cell clustering. However, these graph-based methods cannot rank the importance of the different neighbors for a particular sample in the downstream cancer subtype analyses. In this study, we introduce omicsGAT, a graph attention network (GAT) model to integrate graph-based learning with an attention mechanism for RNA-seq data analysis. The multi-head attention mechanism in omicsGAT can more effectively secure information of a particular sample by assigning different attention coefficients to its neighbors. Comprehensive experiments on The Cancer Genome Atlas (TCGA) breast cancer and bladder cancer bulk RNA-seq data and two single-cell RNA-seq datasets validate that (1) the proposed model can effectively integrate neighborhood information of a sample and learn an embedding vector to improve disease phenotype prediction, cancer patient stratification, and cell clustering of the sample and (2) the attention matrix generated from the multi-head attention coefficients provides more useful information compared to the sample correlation-based adjacency matrix. From the results, we can conclude that some neighbors play a more important role than others in cancer subtype analyses of a particular sample based on the attention coefficient.

Precise prediction of multiple anticancer drug efficacy using multi target regression and support vector regression analysis

BACKGROUND AND OBJECTIVES

The prediction of multiple drug efficacies using machine learning prediction techniques based on clinical and molecular attributes of tumors is a new approach in the field of precision medicine of oncology. The selection of suitable and effective therapeutic drugs among the potential drugs is performed computationally considering the tumor features. In this study, we developed and validated machine learning models to predict the efficacy of five anti-cancer drugs according to the clinical and molecular attributes of 30 oral squamous cell carcinoma (OSCC) cohorts. This sounds a bit odd - consider: Ranking of the drugs was achieved using their apoptotic priming.

METHODS

We developed multiple drug efficacy prediction models based on three types of tumor characteristics by applying machine learning methods, including multi-target regression (MTR) and support vector regression (SVR). The prediction accuracy of existing machine learning methods was enhanced by introducing novel pre-processing techniques to develop Enhanced MTR (E_MTR), Enhanced Log-based MTR (EL_MTR), Enhanced Multi-target SVR (EM_SVR), and Enhanced Log-based Multi-target SVR (ELM_SVR). As a unique capability, ELM_SVR and EL_MTR rank the drugs based on their predicted efficacy. All the drug efficacy prediction models were built using OSCC real samples and theoretical samples. The best model was selected was based on dataset size and evaluation metrics, such as error terms, residuals and parameter tuning, and cross-validated (CV) using 30 real samples and 340 theoretical samples.

RESULTS

When 30 real tumor samples were used for the train-test and CV methods, MTR models predicted the efficacy with less error than SVR models. Comparatively, using 340 theoretical samples for the train-test and CV methods, though MTR improved the performance, SVR predicted the efficacy with zero error. We found that, for small samples, the proposed MTR provided a 0.01 difference between actual apoptotic priming and predicted priming of five drugs. For large samples, the predicted values by the proposed SVR had a difference of 0.00001. The error terms (Actual vs. Predicted) also reveal that the enhanced log model is suitable when MTR is applied. Meanwhile, the enhanced model is suitable for SVR learning for multiple drug efficacy prediction. It was found that the predicted ranks of the drugs based on the multi-targeted efficacy prediction exactly match the actual rankings.

CONCLUSION

We developed efficient statistical and machine learning models using MTR and SVR analysis for anticancer drug efficacy, which will be useful in the field of precision medicine to choose the most suitable drugs in personalized manner. The performance results of the proposed enhanced ranking techniques are described as follows: i) EL_MTR is the best to predict multiple anticancer drug efficacies and improve the accuracy of ranking drugs, irrespective of sample size; and ii) ELM_SVR performs better than other MTR models with a large sample size and precise ranking process.

EpICC: A Bayesian neural network model with uncertainty correction for a more accurate classification of cancer

Accurate classification of cancers into their types and subtypes holds the key for choosing the right treatment strategy and can greatly impact patient well-being. However, existence of large-scale variations in the molecular processes driving even a single type of cancer can make accurate classification a challenging problem. Therefore, improved and robust methods for classification are absolutely critical. Although deep learning-based methods for cancer classification have been proposed earlier, they all provide point estimates for predictions without any measure of confidence and thus, can fall short in real-world applications where key decisions are to be made based on the predictions of the classifier. Here we report a Bayesian neural network-based model for classification of cancer types as well as sub-types from transcriptomic data. This model reported a measure of confidence with each prediction through analysis of epistemic uncertainty. We incorporated an uncertainty correction step with the Bayesian network-based model to greatly enhance prediction accuracy of cancer types (> 97% accuracy) and sub-types (> 80%). Our work suggests that reporting uncertainty measure with each classification can enable more accurate and informed decision-making that can be highly valuable in clinical settings.

AFExNet: An Adversarial Autoencoder for Differentiating Breast Cancer Sub-Types and Extracting Biologically Relevant Genes

Technological advancements in high-throughput genomics enable the generation of complex and large data sets that can be used for classification, clustering, and bio-marker identification. Modern deep learning algorithms provide us with the opportunity of finding most significant features in such huge dataset to characterize diseases (e.g., cancer) and their sub-types. Thus, developing such deep learning method, which can successfully extract meaningful features from various breast cancer sub-types, is of current research interest. In this paper, we develop dual stage (unsupervised pre-training and supervised fine-tuning) neural network architecture termed AFExNet based on adversarial auto-encoder (AAE) to extract features from high dimensional genetic data. We evaluated the performance of our model through twelve different supervised classifiers to verify the usefulness of the new features using public RNA-Seq dataset of breast cancer. AFExNet provides consistent results in all performance metrics across twelve different classifiers which makes our model classifier independent. We also develop a method named 'TopGene' to find highly weighted genes from the latent space which could be useful for finding cancer bio-markers. Put together, AFExNet has great potential for biological data to accurately and effectively extract features. Our work is fully reproducible and source code can be downloaded from Github: https://github.com/NeuroSyd/breast-cancer-sub-types.

GraphChrom: A Novel Graph-Based Framework for Cancer Classification Using Chromosomal Rearrangement Endpoints

Chromosomal rearrangements are generally a consequence of improperly repaired double-strand breaks in DNA. These genomic aberrations can be a driver of cancers. Here, we investigated the use of chromosomal rearrangements for classification of cancer tumors and the effect of inter- and intrachromosomal rearrangements in cancer classification. We used data from the Catalogue of Somatic Mutations in Cancer (COSMIC) for breast, pancreatic, and prostate cancers, for which the COSMIC dataset reports the highest number of chromosomal aberrations. We developed a framework known as GraphChrom for cancer classification. GraphChrom was developed using a graph neural network which models the complex structure of chromosomal aberrations (CA) and provides local connectivity between the aberrations. The proposed framework illustrates three important contributions to the field of cancers. Firstly, it successfully classifies cancer types and subtypes. Secondly, it evolved into a novel data extraction technique which can be used to extract more informative graphs (informative aberrations associated with a sample); and thirdly, it predicts that interCAs (rearrangements between two or more chromosomes) are more effective in cancer prediction than intraCAs (rearrangements within the same chromosome), although intraCAs are three times more likely to occur than intraCAs.

Molecular Subtyping of Cancer Based on Robust Graph Neural Network and Multi-Omics Data Integration

Accurate molecular subtypes prediction of cancer patients is significant for personalized cancer diagnosis and treatments. Large amount of multi-omics data and the advancement of data-driven methods are expected to facilitate molecular subtyping of cancer. Most existing machine learning–based methods usually classify samples according to single omics data, fail to integrate multi-omics data to learn comprehensive representations of the samples, and ignore that information transfer and aggregation among samples can better represent them and ultimately help in classification. We propose a novel framework named multi-omics graph convolutional network (M-GCN) for molecular subtyping based on robust graph convolutional networks integrating multi-omics data. We first apply the Hilbert–Schmidt independence criterion least absolute shrinkage and selection operator (HSIC Lasso) to select the molecular subtype-related transcriptomic features and then construct a sample–sample similarity graph with low noise by using these features. Next, we take the selected gene expression, single nucleotide variants (SNV), and copy number variation (CNV) data as input and learn the multi-view representations of samples. On this basis, a robust variant of graph convolutional network (GCN) model is finally developed to obtain samples’ new representations by aggregating their subgraphs. Experimental results of breast and stomach cancer demonstrate that the classification performance of M-GCN is superior to other existing methods. Moreover, the identified subtype-specific biomarkers are highly consistent with current clinical understanding and promising to assist accurate diagnosis and targeted drug development.

Single-cell sequencing: a promising approach for uncovering the mechanisms of tumor metastasis

Single-cell sequencing (SCS) is an emerging high-throughput technology that can be used to study the genomics, transcriptomics, and epigenetics at a single cell level. SCS is widely used in the diagnosis and treatment of various diseases, including cancer. Over the years, SCS has gradually become an effective clinical tool for the exploration of tumor metastasis mechanisms and the development of treatment strategies. Currently, SCS can be used not only to analyze metastasis-related malignant biological characteristics, such as tumor heterogeneity, drug resistance, and microenvironment, but also to construct metastasis-related cell maps for predicting and monitoring the dynamics of metastasis. SCS is also used to identify therapeutic targets related to metastasis as it provides insights into the distribution of tumor cell subsets and gene expression differences between primary and metastatic tumors. Additionally, SCS techniques in combination with artificial intelligence (AI) are used in liquid biopsy to identify circulating tumor cells (CTCs), thereby providing a novel strategy for treating tumor metastasis. In this review, we summarize the potential applications of SCS in the field of tumor metastasis and discuss the prospects and limitations of SCS to provide a theoretical basis for finding therapeutic targets and mechanisms of metastasis.

Multi-Size Deep Learning Based Preoperative Computed Tomography Signature for Prognosis Prediction of Colorectal Cancer

Background: Preoperative and postoperative evaluation of colorectal cancer (CRC) patients is crucial for subsequent treatment guidance. Our study aims to provide a timely and rapid assessment of the prognosis of CRC patients with deep learning according to non-invasive preoperative computed tomography (CT) and explore the underlying biological explanations.

Methods: A total of 808 CRC patients with preoperative CT (development cohort: n = 426, validation cohort: n = 382) were enrolled in our study. We proposed a novel end-to-end Multi-Size Convolutional Neural Network (MSCNN) to predict the risk of CRC recurrence with CT images (CT signature). The prognostic performance of CT signature was evaluated by Kaplan-Meier curve. An integrated nomogram was constructed to improve the clinical utility of CT signature by combining with other clinicopathologic factors. Further visualization and correlation analysis for CT deep features with paired gene expression profiles were performed to reveal the molecular characteristics of CRC tumors learned by MSCNN in radiographic imaging.

Results: The Kaplan-Meier analysis showed that CT signature was a significant prognostic factor for CRC disease-free survival (DFS) prediction [development cohort: hazard ratio (HR): 50.7, 95% CI: 28.4–90.6, p &lt; 0.001; validation cohort: HR: 2.04, 95% CI: 1.44–2.89, p &lt; 0.001]. Multivariable analysis confirmed the independence prognostic value of CT signature (development cohort: HR: 30.7, 95% CI: 19.8–69.3, p &lt; 0.001; validation cohort: HR: 1.83, 95% CI: 1.19–2.83, p = 0.006). Dimension reduction and visualization of CT deep features demonstrated a high correlation with the prognosis of CRC patients. Functional pathway analysis further indicated that CRC patients with high CT signature presented down-regulation of several immunology pathways. Correlation analysis found that CT deep features were mainly associated with activation of metabolic and proliferative pathways.

Conclusions: Our deep learning based preoperative CT signature can effectively predict prognosis of CRC patients. Integration analysis of multi-omic data revealed that some molecular characteristics of CRC tumor can be captured by deep learning in CT images.