Global proteomics profiling improves drug sensitivity prediction: results from a multi-omics, pan-cancer modeling approach

Methods for multi-omic data integration in cancer research

Multi-omics data integration is a term that refers to the process of combining and analyzing data from different omic experimental sources, such as genomics, transcriptomics, methylation assays, and microRNA sequencing, among others. Such data integration approaches have the potential to provide a more comprehensive functional understanding of biological systems and has numerous applications in areas such as disease diagnosis, prognosis and therapy. However, quantitative integration of multi-omic data is a complex task that requires the use of highly specialized methods and approaches. Here, we discuss a number of data integration methods that have been developed with multi-omics data in view, including statistical methods, machine learning approaches, and network-based approaches. We also discuss the challenges and limitations of such methods and provide examples of their applications in the literature. Overall, this review aims to provide an overview of the current state of the field and highlight potential directions for future research.

Multi-OMICS approaches in cancer biology: New era in cancer therapy

Innovative multi-omics frameworks integrate diverse datasets from the same patients to enhance our understanding of the molecular and clinical aspects of cancers. Advanced omics and multi-view clustering algorithms present unprecedented opportunities for classifying cancers into subtypes, refining survival predictions and treatment outcomes, and unravelling key pathophysiological processes across various molecular layers. However, with the increasing availability of cost-effective high-throughput technologies (HTT) that generate vast amounts of data, analyzing single layers often falls short of establishing causal relations. Integrating multi-omics data spanning genomes, epigenomes, transcriptomes, proteomes, metabolomes, and microbiomes offers unique prospects to comprehend the underlying biology of complex diseases like cancer. This discussion explores algorithmic frameworks designed to uncover cancer subtypes, disease mechanisms, and methods for identifying pivotal genomic alterations. It also underscores the significance of multi-omics in tumor classifications, diagnostics, and prognostications. Despite its unparalleled advantages, the integration of multi-omics data has been slow to find its way into everyday clinics. A major hurdle is the uneven maturity of different omics approaches and the widening gap between the generation of large datasets and the capacity to process this data. Initiatives promoting the standardization of sample processing and analytical pipelines, as well as multidisciplinary training for experts in data analysis and interpretation, are crucial for translating theoretical findings into practical applications.

Mapping the proteogenomic landscape enables prediction of drug response in acute myeloid leukemia

Acute myeloid leukemia is a poor-prognosis cancer commonly stratified by genetic aberrations, but these mutations are often heterogeneous and fail to consistently predict therapeutic response. Here, we combine transcriptomic, proteomic, and phosphoproteomic datasets with ex vivo drug sensitivity data to help understand the underlying pathophysiology of AML beyond mutations. We measure the proteome and phosphoproteome of 210 patients and combine them with genomic and transcriptomic measurements to identify four proteogenomic subtypes that complement existing genetic subtypes. We build a predictor to classify samples into subtypes and map them to a "landscape" that identifies specific drug response patterns. We then build a drug response prediction model to identify drugs that target distinct subtypes and validate our findings on cell lines representing various stages of quizartinib resistance. Our results show how multiomics data together with drug sensitivity data can inform therapy stratification and drug combinations in AML.

MMCL-CDR: enhancing cancer drug response prediction with multi-omics and morphology images contrastive representation learning

MOTIVATION

Cancer is a complex disease that results in a significant number of global fatalities. Treatment strategies can vary among patients, even if they have the same type of cancer. The application of precision medicine in cancer shows promise for treating different types of cancer, reducing healthcare expenses, and improving recovery rates. To achieve personalized cancer treatment, machine learning models have been developed to predict drug responses based on tumor and drug characteristics. However, current studies either focus on constructing homogeneous networks from single data source or heterogeneous networks from multiomics data. While multiomics data have shown potential in predicting drug responses in cancer cell lines, there is still a lack of research that effectively utilizes insights from different modalities. Furthermore, effectively utilizing the multimodal knowledge of cancer cell lines poses a challenge due to the heterogeneity inherent in these modalities.

RESULTS

To address these challenges, we introduce MMCL-CDR (Multimodal Contrastive Learning for Cancer Drug Responses), a multimodal approach for cancer drug response prediction that integrates copy number variation, gene expression, morphology images of cell lines, and chemical structure of drugs. The objective of MMCL-CDR is to align cancer cell lines across different data modalities by learning cell line representations from omic and image data, and combined with structural drug representations to enhance the prediction of cancer drug responses (CDR). We have carried out comprehensive experiments and show that our model significantly outperforms other state-of-the-art methods in CDR prediction. The experimental results also prove that the model can learn more accurate cell line representation by integrating multiomics and morphological data from cell lines, thereby improving the accuracy of CDR prediction. In addition, the ablation study and qualitative analysis also confirm the effectiveness of each part of our proposed model. Last but not least, MMCL-CDR opens up a new dimension for cancer drug response prediction through multimodal contrastive learning, pioneering a novel approach that integrates multiomics and multimodal drug and cell line modeling.

AVAILABILITY AND IMPLEMENTATION

MMCL-CDR is available at https://github.com/catly/MMCL-CDR.

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.

Realm of proteomics in breast cancer management and drug repurposing to alleviate intricacies of treatment

Breast cancer, a multi-networking heterogeneous disease, has emerged as a serious impediment to progress in clinical oncology. Although technological advancements and emerging cancer research studies have mitigated breast cancer lethality, a precision cancer-oriented solution has not been achieved. Thus, this review will persuade the acquiescence of proteomics-based diagnostic and therapeutic options in breast cancer management. Recently, the evidence of breast cancer health surveillance through imaging proteomics, single-cell proteomics, interactomics, and post-translational modification (PTM) tracking, to construct proteome maps and proteotyping for stage-specific and sample-specific cancer subtyping have outperformed conventional ways of dealing with breast cancer by increasing diagnostic efficiency, prognostic value, and predictive response. Additionally, the paradigm shift in applied proteomics for designing a chemotherapy regimen to identify novel drug targets with minor adverse effects has been elaborated. Finally, the potential of proteomics in alleviating the occurrence of chemoresistance and enhancing reprofiled drugs' effectiveness to combat therapeutic obstacles has been discussed. Owing to the enormous potential of proteomics techniques, the clinical recognition of proteomics in breast cancer management can be achievable and therapeutic intricacies can be surmountable.

The proteomic landscape of soft tissue sarcomas

Soft tissue sarcomas (STS) are rare and diverse mesenchymal cancers with limited treatment options. Here we undertake comprehensive proteomic profiling of tumour specimens from 321 STS patients representing 11 histological subtypes. Within leiomyosarcomas, we identify three proteomic subtypes with distinct myogenesis and immune features, anatomical site distribution and survival outcomes. Characterisation of undifferentiated pleomorphic sarcomas and dedifferentiated liposarcomas with low infiltrating CD3 + T-lymphocyte levels nominates the complement cascade as a candidate immunotherapeutic target. Comparative analysis of proteomic and transcriptomic profiles highlights the proteomic-specific features for optimal risk stratification in angiosarcomas. Finally, we define functional signatures termed Sarcoma Proteomic Modules which transcend histological subtype classification and show that a vesicle transport protein signature is an independent prognostic factor for distant metastasis. Our study highlights the utility of proteomics for identifying molecular subgroups with implications for risk stratification and therapy selection and provides a rich resource for future sarcoma research.

IGPred-HDnet: Prediction of Immunoglobulin Proteins Using Graphical Features and the Hierarchal Deep Learning-Based Approach

Motivation. Immunoglobulin proteins (IGP) (also called antibodies) are glycoproteins that act as B-cell receptors against external or internal antigens like viruses and bacteria. IGPs play a significant role in diverse cellular processes ranging from adhesion to cell recognition. IGP identifications via the in-silico approach are faster and more cost-effective than wet-lab technological methods. Methods. In this study, we developed an intelligent theoretical deep learning framework, "IGPred-HDnet" for the discrimination of IGPs and non-IGPs. Three types of promising descriptors are feature extraction based on graphical and statistical features (FEGS), amphiphilic pseudo-amino acid composition (Amp-PseAAC), and dipeptide composition (DPC) to extract the graphical, physicochemical, and sequential features. Next, the extracted attributes are evaluated through machine learning, i.e., decision tree (DT), support vector machine (SVM), k-nearest neighbour (KNN), and hierarchical deep network (HDnet) classifiers. The proposed predictor IGPred-HDnet was trained and tested using a 10-fold cross-validation and independent test. Results and Conclusion. The success rates in terms of accuracy (ACC) and Matthew's correlation coefficient (MCC) of IGPred-HDnet on training and independent dataset (Dtrain Dtest) are ACC = 98.00%, 99.10%, and MCC = 0.958, and 0.980 points, respectively. The empirical outcomes demonstrate that the IGPred-HDnet model efficacy on both datasets using the novel FEGS feature and HDnet algorithm achieved superior predictions to other existing computational models. We hope this research will provide great insights into the large-scale identification of IGPs and pharmaceutical companies in new drug design.

Research progress in protein microarrays: Focussing on cancer research

Although several effective treatment modalities have been developed for cancers, the morbidity and mortality associated with cancer continues to increase every year. As one of the most exciting emerging technologies, protein microarrays represent a powerful tool in the field of cancer research because of their advantages such as high throughput, small sample usage, more flexibility, high sensitivity and direct readout of results. In this review, we focus on the research progress in four types of protein microarrays (proteome microarray, antibody microarray, lectin microarray and reversed protein array) with emphasis on their application in cancer research. Finally, we discuss the current challenges faced by protein microarrays and directions for future developments. We firmly believe that this novel systems biology research tool holds immense potential in cancer research and will become an irreplaceable tool in this field.

Transforming Diagnosis and Therapeutics Using Cancer Genomics

In past quarter of the century, much has been understood about the genetic variation and abnormal genes that activate cancer in humans. All the cancers somehow possess alterations in the DNA sequence of cancer cell's genome. In present, we are heading toward the era where it is possible to obtain complete genome of the cancer cells for their better diagnosis, categorization and to explore treatment options.

Artificial intelligence-based multi-omics analysis fuels cancer precision medicine

With biotechnological advancements, innovative omics technologies are constantly emerging that have enabled researchers to access multi-layer information from the genome, epigenome, transcriptome, proteome, metabolome, and more. A wealth of omics technologies, including bulk and single-cell omics approaches, have empowered to characterize different molecular layers at unprecedented scale and resolution, providing a holistic view of tumor behavior. Multi-omics analysis allows systematic interrogation of various molecular information at each biological layer while posing tricky challenges regarding how to extract valuable insights from the exponentially increasing amount of multi-omics data. Therefore, efficient algorithms are needed to reduce the dimensionality of the data while simultaneously dissecting the mysteries behind the complex biological processes of cancer. Artificial intelligence has demonstrated the ability to analyze complementary multi-modal data streams within the oncology realm. The coincident development of multi-omics technologies and artificial intelligence algorithms has fuelled the development of cancer precision medicine. Here, we present state-of-the-art omics technologies and outline a roadmap of multi-omics integration analysis using an artificial intelligence strategy. The advances made using artificial intelligence-based multi-omics approaches are described, especially concerning early cancer screening, diagnosis, response assessment, and prognosis prediction. Finally, we discuss the challenges faced in multi-omics analysis, along with tentative future trends in this field. With the increasing application of artificial intelligence in multi-omics analysis, we anticipate a shifting paradigm in precision medicine becoming driven by artificial intelligence-based multi-omics technologies.

Mutational analysis and protein profiling predict drug sensitivity in multiple myeloma cell lines

INTRODUCTION

Multiple myeloma (MM) is a heterogeneous disease where cancer-driver mutations and aberrant signaling may lead to disease progression and drug resistance. Drug responses vary greatly, and there is an unmet need for biomarkers that can guide precision cancer medicine in this disease.

METHODS

To identify potential predictors of drug sensitivity, we applied integrated data from drug sensitivity screening, mutational analysis and functional signaling pathway profiling in 9 cell line models of MM. We studied the sensitivity to 33 targeted drugs and their association with the mutational status of cancer-driver genes and activity level of signaling proteins.

RESULTS

We found that sensitivity to mitogen-activated protein kinase kinase 1 (MEK1) and phosphatidylinositol-3 kinase (PI3K) inhibitors correlated with mutations in NRAS/KRAS, and PI3K family genes, respectively. Phosphorylation status of MEK1 and protein kinase B (AKT) correlated with sensitivity to MEK and PI3K inhibition, respectively. In addition, we found that enhanced phosphorylation of proteins, including Tank-binding kinase 1 (TBK1), as well as high expression of B cell lymphoma 2 (Bcl-2), correlated with low sensitivity to MEK inhibitors.

DISCUSSION

Taken together, this study shows that mutational status and signaling protein profiling might be used in further studies to predict drug sensitivities and identify resistance markers in MM.

Deep learning and multi-omics approach to predict drug responses in cancer

BACKGROUND

Cancers are genetically heterogeneous, so anticancer drugs show varying degrees of effectiveness on patients due to their differing genetic profiles. Knowing patient's responses to numerous cancer drugs are needed for personalized treatment for cancer. By using molecular profiles of cancer cell lines available from Cancer Cell Line Encyclopedia (CCLE) and anticancer drug responses available in the Genomics of Drug Sensitivity in Cancer (GDSC), we will build computational models to predict anticancer drug responses from molecular features.

RESULTS

We propose a novel deep neural network model that integrates multi-omics data available as gene expressions, copy number variations, gene mutations, reverse phase protein array expressions, and metabolomics expressions, in order to predict cellular responses to known anti-cancer drugs. We employ a novel graph embedding layer that incorporates interactome data as prior information for prediction. Moreover, we propose a novel attention layer that effectively combines different omics features, taking their interactions into account. The network outperformed feedforward neural networks and reported 0.90 for [Formula: see text] values for prediction of drug responses from cancer cell lines data available in CCLE and GDSC.

CONCLUSION

The outstanding results of our experiments demonstrate that the proposed method is capable of capturing the interactions of genes and proteins, and integrating multi-omics features effectively. Furthermore, both the results of ablation studies and the investigations of the attention layer imply that gene mutation has a greater influence on the prediction of drug responses than other omics data types. Therefore, we conclude that our approach can not only predict the anti-cancer drug response precisely but also provides insights into reaction mechanisms of cancer cell lines and drugs as well.

Proteomic and phosphoproteomic measurements enhance ability to predict ex vivo drug response in AML

Acute Myeloid Leukemia (AML) affects 20,000 patients in the US annually with a five-year survival rate of approximately 25%. One reason for the low survival rate is the high prevalence of clonal evolution that gives rise to heterogeneous sub-populations of leukemic cells with diverse mutation spectra, which eventually leads to disease relapse. This genetic heterogeneity drives the activation of complex signaling pathways that is reflected at the protein level. This diversity makes it difficult to treat AML with targeted therapy, requiring custom patient treatment protocols tailored to each individual's leukemia. Toward this end, the Beat AML research program prospectively collected genomic and transcriptomic data from over 1000 AML patients and carried out ex vivo drug sensitivity assays to identify genomic signatures that could predict patient-specific drug responses. However, there are inherent weaknesses in using only genetic and transcriptomic measurements as surrogates of drug response, particularly the absence of direct information about phosphorylation-mediated signal transduction. As a member of the Clinical Proteomic Tumor Analysis Consortium, we have extended the molecular characterization of this cohort by collecting proteomic and phosphoproteomic measurements from a subset of these patient samples (38 in total) to evaluate the hypothesis that proteomic signatures can improve the ability to predict response to 26 drugs in AML ex vivo samples. In this work we describe our systematic, multi-omic approach to evaluate proteomic signatures of drug response and compare protein levels to other markers of drug response such as mutational patterns. We explore the nuances of this approach using two drugs that target key pathways activated in AML: quizartinib (FLT3) and trametinib (Ras/MEK), and show how patient-derived signatures can be interpreted biologically and validated in cell lines. In conclusion, this pilot study demonstrates strong promise for proteomics-based patient stratification to assess drug sensitivity in AML.

Tissue-specific identification of multi-omics features for pan-cancer drug response prediction

Current statistical models for drug response prediction and biomarker identification fall short in leveraging the shared and unique information from various cancer tissues and multi-omics profiles. We developed mix-lasso model that introduces an additional sample group penalty term to capture tissue-specific effects of features on pan-cancer response prediction. The mix-lasso model takes into account both the similarity between drug responses (i.e., multi-task learning), and the heterogeneity between multi-omics data (multi-modal learning). When applied to large-scale pharmacogenomics dataset from Cancer Therapeutics Response Portal, mix-lasso enabled accurate drug response predictions and identification of tissue-specific predictive features in the presence of various degrees of missing data, drug-drug correlations, and high-dimensional and correlated genomic and molecular features that often hinder the use of statistical approaches in drug response modeling. Compared to tree lasso model, mix-lasso identified a smaller number of tissue-specific features, hence making the model more interpretable and stable for drug discovery applications.

•

Pan-cancer cell lines provide a test bench for exploring gene-drug relationships

•

Multi-omics data were integrated with pharmacological profiles for joint modeling

•

Mix-lasso identifies tissue-specific biomarkers predictive of multi-drug responses

•

Mix-lasso provides small number of stable features for drug discovery applications

Application of informatics in cancer research and clinical practice: Opportunities and challenges

Cancer informatics has significantly progressed in the big data era. We summarize the application of informatics approaches to the cancer domain from both the informatics perspective (e.g., data management and data science) and the clinical perspective (e.g., cancer screening, risk assessment, diagnosis, treatment, and prognosis). We discuss various informatics methods and tools that are widely applied in cancer research and practices, such as cancer databases, data standards, terminologies, high-throughput omics data mining, machine-learning algorithms, artificial intelligence imaging, and intelligent radiation. We also address the informatics challenges within the cancer field that pursue better treatment decisions and patient outcomes, and focus on how informatics can provide opportunities for cancer research and practices. Finally, we conclude that the interdisciplinary nature of cancer informatics and collaborations are major drivers for future research and applications in clinical practices. It is hoped that this review is instrumental for cancer researchers and clinicians with its informatics-specific insights.

Integrative multi-omics and drug response profiling of childhood acute lymphoblastic leukemia cell lines

Acute lymphoblastic leukemia (ALL) is the most common childhood cancer. Although standard-of-care chemotherapeutics are sufficient for most ALL cases, there are subsets of patients with poor response who relapse in disease. The biology underlying differences between subtypes and their response to therapy has only partially been explained by genetic and transcriptomic profiling. Here, we perform comprehensive multi-omic analyses of 49 readily available childhood ALL cell lines, using proteomics, transcriptomics, and pharmacoproteomic characterization. We connect the molecular phenotypes with drug responses to 528 oncology drugs, identifying drug correlations as well as lineage-dependent correlations. We also identify the diacylglycerol-analog bryostatin-1 as a therapeutic candidate in the MEF2D-HNRNPUL1 fusion high-risk subtype, for which this drug activates pro-apoptotic ERK signaling associated with molecular mediators of pre-B cell negative selection. Our data is the foundation for the interactive online Functional Omics Resource of ALL (FORALL) with navigable proteomics, transcriptomics, and drug sensitivity profiles at https://proteomics.se/forall.

Childhood acute lymphoblastic leukemia is characterised by a range of genetic aberrations. Here, the authors use multi-omics profiling of ALL cell lines to connect molecular phenotypes and drug responses to provide an interactive resource of drug sensitivity.

Hyaluronic Acid Nanoparticles for Immunogenic Chemotherapy of Leukemia and T-Cell Lymphoma

Ratiometric delivery of combination chemotherapy can achieve therapeutic efficacy based on synergistic interactions between drugs. It is critical to design such combinations with drugs that complement each other and reduce cancer growth through multiple mechanisms. Using hyaluronic acid (HA) as a carrier, two chemotherapeutic agents-doxorubicin (DOX) and camptothecin (CPT)-were incorporated and tested for their synergistic potency against a broad panel of blood-cancer cell lines. The pair also demonstrated the ability to achieve immunogenic cell death by increasing the surface exposure levels of Calreticulin, thereby highlighting its ability to induce apoptosis via an alternate pathway. Global proteomic profiling of cancer cells treated with HA-DOX-CPT identified pathways that could potentially predict patient sensitivity to HA-DOX-CPT. This lays the foundation for further exploration of integrating drug delivery and proteomics in personalized immunogenic chemotherapy.

Predicting drug sensitivity of cancer cells based on DNA methylation levels

Cancer cell lines, which are cell cultures derived from tumor samples, represent one of the least expensive and most studied preclinical models for drug development. Accurately predicting drug responses for a given cell line based on molecular features may help to optimize drug-development pipelines and explain mechanisms behind treatment responses. In this study, we focus on DNA methylation profiles as one type of molecular feature that is known to drive tumorigenesis and modulate treatment responses. Using genome-wide, DNA methylation profiles from 987 cell lines in the Genomics of Drug Sensitivity in Cancer database, we used machine-learning algorithms to evaluate the potential to predict cytotoxic responses for eight anti-cancer drugs. We compared the performance of five classification algorithms and four regression algorithms representing diverse methodologies, including tree-, probability-, kernel-, ensemble-, and distance-based approaches. We artificially subsampled the data to varying degrees, aiming to understand whether training based on relatively extreme outcomes would yield improved performance. When using classification or regression algorithms to predict discrete or continuous responses, respectively, we consistently observed excellent predictive performance when the training and test sets consisted of cell-line data. Classification algorithms performed best when we trained the models using cell lines with relatively extreme drug-response values, attaining area-under-the-receiver-operating-characteristic-curve values as high as 0.97. The regression algorithms performed best when we trained the models using the full range of drug-response values, although this depended on the performance metrics we used. Finally, we used patient data from The Cancer Genome Atlas to evaluate the feasibility of classifying clinical responses for human tumors based on models derived from cell lines. Generally, the algorithms were unable to identify patterns that predicted patient responses reliably; however, predictions by the Random Forests algorithm were significantly correlated with Temozolomide responses for low-grade gliomas.

MicroRNA Interactome Multiomics Characterization for Cancer Research and Personalized Medicine: An Expert Review

MicroRNAs (miRNAs) that are mutually modulated by their interacting partners (interactome) are being increasingly noted for their significant role in pathogenesis and treatment of various human cancers. Recently, miRNA interactome dissected with multiomics approaches has been the subject of focus since individual tools or methods failed to provide the necessary comprehensive clues on the complete interactome. Even though single-omics technologies such as proteomics can uncover part of the interactome, the biological and clinical understanding still remain incomplete. In this study, we present an expert review of studies involving multiomics approaches to identification of miRNA interactome and its application in mechanistic characterization, classification, and therapeutic target identification in a variety of cancers, and with a focus on proteomics. We also discuss individual or multiple miRNA-based interactome identification in various pathological conditions of relevance to clinical medicine. Various new single-omics methods that can be integrated into multiomics cancer research and the computational approaches to analyze and predict miRNA interactome are also highlighted in this review. In all, we contextulize the power of multiomics approaches and the importance of the miRNA interactome to achieve the vision and practice of predictive, preventive, and personalized medicine in cancer research and clinical oncology.

A gentle introduction to understanding preclinical data for cancer pharmaco-omic modeling

A central goal of precision oncology is to administer an optimal drug treatment to each cancer patient. A common preclinical approach to tackle this problem has been to characterize the tumors of patients at the molecular and drug response levels, and employ the resulting datasets for predictive in silico modeling (mostly using machine learning). Understanding how and why the different variants of these datasets are generated is an important component of this process. This review focuses on providing such introduction aimed at scientists with little previous exposure to this research area.

Extensive three-dimensional intratumor proteomic heterogeneity revealed by multiregion sampling in high-grade serous ovarian tumor specimens

Enriched tumor epithelium, tumor-associated stroma, and whole tissue were collected by laser microdissection from thin sections across spatially separated levels of ten high-grade serous ovarian carcinomas (HGSOCs) and analyzed by mass spectrometry, reverse phase protein arrays, and RNA sequencing. Unsupervised analyses of protein abundance data revealed independent clustering of an enriched stroma and enriched tumor epithelium, with whole tumor tissue clustering driven by overall tumor "purity." Comparing these data to previously defined prognostic HGSOC molecular subtypes revealed protein and transcript expression from tumor epithelium correlated with the differentiated subtype, whereas stromal proteins (and transcripts) correlated with the mesenchymal subtype. Protein and transcript abundance in the tumor epithelium and stroma exhibited decreased correlation in samples collected just hundreds of microns apart. These data reveal substantial tumor microenvironment protein heterogeneity that directly bears on prognostic signatures, biomarker discovery, and cancer pathophysiology and underscore the need to enrich cellular subpopulations for expression profiling.

Machine learning in the prediction of cancer therapy

Resistance to therapy remains a major cause of cancer treatment failures, resulting in many cancer-related deaths. Resistance can occur at any time during the treatment, even at the beginning. The current treatment plan is dependent mainly on cancer subtypes and the presence of genetic mutations. Evidently, the presence of a genetic mutation does not always predict the therapeutic response and can vary for different cancer subtypes. Therefore, there is an unmet need for predictive models to match a cancer patient with a specific drug or drug combination. Recent advancements in predictive models using artificial intelligence have shown great promise in preclinical settings. However, despite massive improvements in computational power, building clinically useable models remains challenging due to a lack of clinically meaningful pharmacogenomic data. In this review, we provide an overview of recent advancements in therapeutic response prediction using machine learning, which is the most widely used branch of artificial intelligence. We describe the basics of machine learning algorithms, illustrate their use, and highlight the current challenges in therapy response prediction for clinical practice.

Machine learning analyses of antibody somatic mutations predict immunoglobulin light chain toxicity

In systemic light chain amyloidosis (AL), pathogenic monoclonal immunoglobulin light chains (LC) form toxic aggregates and amyloid fibrils in target organs. Prompt diagnosis is crucial to avoid permanent organ damage, but delayed diagnosis is common because symptoms usually appear only after strong organ involvement. Here we present LICTOR, a machine learning approach predicting LC toxicity in AL, based on the distribution of somatic mutations acquired during clonal selection. LICTOR achieves a specificity and a sensitivity of 0.82 and 0.76, respectively, with an area under the receiver operating characteristic curve (AUC) of 0.87. Tested on an independent set of 12 LCs sequences with known clinical phenotypes, LICTOR achieves a prediction accuracy of 83%. Furthermore, we are able to abolish the toxic phenotype of an LC by in silico reverting two germline-specific somatic mutations identified by LICTOR, and by experimentally assessing the loss of in vivo toxicity in a Caenorhabditis elegans model. Therefore, LICTOR represents a promising strategy for AL diagnosis and reducing high mortality rates in AL.

Precision Oncology, Signaling and Anticancer Agents in Cancer Therapeutics

BACKGROUND

The global alliance for genomics and healthcare facilities provides innovational solutions to expedite research and clinical practices for complex and incurable health conditions. Precision oncology is an emerging field explicitly tailored to facilitate cancer diagnosis, prevention and treatment based on patients' genetic profile. Advancements in "omics" techniques, next-generation sequencing, artificial intelligence and clinical trial designs provide a platform for assessing the efficacy and safety of combination therapies and diagnostic procedures.

METHOD

Data were collected from Pubmed and Google scholar using keywords: "Precision medicine", "precision medicine and cancer", "anticancer agents in precision medicine" and reviewed comprehensively.

RESULTS

Personalized therapeutics including immunotherapy, cancer vaccines, serve as a groundbreaking solution for cancer treatment. Herein, we take a measurable view of precision therapies and novel diagnostic approaches targeting cancer treatment. The contemporary applications of precision medicine have also been described along with various hurdles identified in the successful establishment of precision therapeutics.

CONCLUSION

This review highlights the key breakthroughs related to immunotherapies, targeted anticancer agents, and target interventions related to cancer signaling mechanisms. The success story of this field in context to drug resistance, safety, patient survival and in improving quality of life is yet to be elucidated. We conclude that, in the near future, the field of individualized treatments may truly revolutionize the nature of cancer patient care.

Impact of between-tissue differences on pan-cancer predictions of drug sensitivity

Increased availability of drug response and genomics data for many tumor cell lines has accelerated the development of pan-cancer prediction models of drug response. However, it is unclear how much between-tissue differences in drug response and molecular characteristics may contribute to pan-cancer predictions. Also unknown is whether the performance of pan-cancer models could vary by cancer type. Here, we built a series of pan-cancer models using two datasets containing 346 and 504 cell lines, each with MEK inhibitor (MEKi) response and mRNA expression, point mutation, and copy number variation data, and found that, while the tissue-level drug responses are accurately predicted (between-tissue ρ = 0.88–0.98), only 5 of 10 cancer types showed successful within-tissue prediction performance (within-tissue ρ = 0.11–0.64). Between-tissue differences make substantial contributions to the performance of pan-cancer MEKi response predictions, as exclusion of between-tissue signals leads to a decrease in Spearman’s ρ from a range of 0.43–0.62 to 0.30–0.51. In practice, joint analysis of multiple cancer types usually has a larger sample size, hence greater power, than for one cancer type; and we observe that higher accuracy of pan-cancer prediction of MEKi response is almost entirely due to the sample size advantage. Success of pan-cancer prediction reveals how drug response in different cancers may invoke shared regulatory mechanisms despite tissue-specific routes of oncogenesis, yet predictions in different cancer types require flexible incorporation of between-cancer and within-cancer signals. As most datasets in genome sciences contain multiple levels of heterogeneity, careful parsing of group characteristics and within-group, individual variation is essential when making robust inference.

One of the central goals for precision oncology is to tailor treatment of individual tumors by their molecular characteristics. While drug response predictions have traditionally been sought within each cancer type, it has long been hoped to develop more robust predictions by jointly considering diverse cancer types. While such pan-cancer approaches have improved in recent years, it remains unclear whether between-tissue differences are contributing to the reported pan-cancer prediction performance. This concern stems from the observation that, when cancer types differ in both molecular features and drug response, strong predictive information can come mainly from differences among tissue types. Our study finds that both between- and within-cancer type signals provide substantial contributions to pan-cancer drug response prediction models, and about half of the cancer types examined are poorly predicted despite strong overall performance across all cancer types. We also find that pan-cancer prediction models perform similarly or better than cancer type-specific models, and in many cases the advantage of pan-cancer models is due to the larger number of samples available for pan-cancer analysis. Our results highlight tissue-of-origin as a key consideration for pan-cancer drug response prediction models, and recommend cancer type-specific considerations when translating pan-cancer prediction models for clinical use.

Protein Signature Predicts Response to Neoadjuvant Treatment With Chemotherapy and Bevacizumab in HER2-Negative Breast Cancers

PURPOSE

Antiangiogenic therapy using bevacizumab has proven effective for a number of cancers; however, in breast cancer (BC), there is an unmet need to identify patients who benefit from such treatment.

PATIENTS AND METHODS

In the NeoAva phase II clinical trial, patients (N = 132) with large (≥ 25 mm) human epidermal growth factor receptor 2 (HER2)-negative primary tumors were randomly assigned 1:1 to treatment with neoadjuvant chemotherapy (CTx) alone or in combination with bevacizumab (Bev plus CTx). The ratio of the tumor size after relative to before treatment was calculated into a continuous response scale. Tumor biopsies taken prior to neoadjuvant treatment were analyzed by reverse-phase protein arrays (RPPA) for expression levels of 210 BC-relevant (phospho-) proteins. Lasso regression was used to derive a predictor of tumor shrinkage from the expression of selected proteins prior to treatment.

RESULTS

We identified a nine-protein signature score named vascular endothelial growth factor inhibition response predictor (ViRP) for use in the Bev plus CTx treatment arm able to predict with accuracy pathologic complete response (pCR) (area under the curve [AUC] = 0.85; 95% CI, 0.74 to 0.97) and low residual cancer burden (RCB 0/I) (AUC = 0.80; 95% CI, 0.68 to 0.93). The ViRP score was significantly lower in patients with pCR (P < .001) and in patients with low RCB (P < .001). The ViRP score was internally validated on mRNA data and the resultant surrogate mRNA ViRP score significantly separated the pCR patients (P = .016). Similarly, the mRNA ViRP score was validated (P < .001) in an independent phase II clinical trial (PROMIX).

CONCLUSION

Our ViRP score, integrating the expression of nine proteins and validated on mRNA data both internally and in an independent clinical trial, may be used to increase the likelihood of benefit from treatment with bevacizumab combined with chemotherapy in patients with HER2-negative BC.

Artificial Intelligence (AI)-Based Systems Biology Approaches in Multi-Omics Data Analysis of Cancer

Cancer is the manifestation of abnormalities of different physiological processes involving genes, DNAs, RNAs, proteins, and other biomolecules whose profiles are reflected in different omics data types. As these bio-entities are very much correlated, integrative analysis of different types of omics data, multi-omics data, is required to understanding the disease from the tumorigenesis to the disease progression. Artificial intelligence (AI), specifically machine learning algorithms, has the ability to make decisive interpretation of "big"-sized complex data and, hence, appears as the most effective tool for the analysis and understanding of multi-omics data for patient-specific observations. In this review, we have discussed about the recent outcomes of employing AI in multi-omics data analysis of different types of cancer. Based on the research trends and significance in patient treatment, we have primarily focused on the AI-based analysis for determining cancer subtypes, disease prognosis, and therapeutic targets. We have also discussed about AI analysis of some non-canonical types of omics data as they have the capability of playing the determiner role in cancer patient care. Additionally, we have briefly discussed about the data repositories because of their pivotal role in multi-omics data storing, processing, and analysis.

Patient-specific multi-omics models and the application in personalized combination therapy

The rapid advancement of high-throughput technologies and sharp decrease in cost have opened up the possibility to generate large amount of multi-omics data on an individual basis. The development of high-throughput -omics, including genomics, epigenomics, transcriptomics, proteomics, metabolomics and microbiomics, enables the application of multi-omics technologies in the clinical settings. Combination therapy, defined as disease treatment with two or more drugs to achieve efficacy with lower doses or lower drug toxicity, is the basis for the care of diseases like cancer. Patient-specific multi-omics data integration can help the identification and development of combination therapies. In this review, we provide an overview of different -omics platforms, and discuss the methods for multi-omics, high-throughput, data integration, personalized combination therapy.

Integrative analysis of DNA methylation and gene expression in papillary renal cell carcinoma

Patterns of DNA methylation are significantly altered in cancers. Interpreting the functional consequences of DNA methylation requires the integration of multiple forms of data. The recent advancement in the next-generation sequencing can help to decode this relationship and in biomarker discovery. In this study, we investigated the methylation patterns of papillary renal cell carcinoma (PRCC) and its relationship with the gene expression using The Cancer Genome Atlas (TCGA) multi-omics data. We found that the promoter and body of tumor suppressor genes, microRNAs and gene clusters and families, including cadherins, protocadherins, claudins and collagens, are hypermethylated in PRCC. Hypomethylated genes in PRCC are associated with the immune function. The gene expression of several novel candidate genes, including interleukin receptor IL17RE and immune checkpoint genes HHLA2, SIRPA and HAVCR2, shows a significant correlation with DNA methylation. We also developed machine learning models using features extracted from single and multi-omics data to distinguish early and late stages of PRCC. A comparative study of different feature selection algorithms, predictive models, data integration techniques and representations of methylation data was performed. Integration of both gene expression and DNA methylation features improved the performance of models in distinguishing tumor stages. In summary, our study identifies PRCC driver genes and proposes predictive models based on both DNA methylation and gene expression. These results on PRCC will aid in targeted experiments and provide a strategy to improve the classification accuracy of tumor stages.

Rethinking Drug Repositioning and Development with Artificial Intelligence, Machine Learning, and Omics

Pharmaceutical industry and the art and science of drug development are sorely in need of novel transformative technologies in the current age of digital health and artificial intelligence (AI). Often described as game-changing technologies, AI and machine learning algorithms have slowly but surely begun to revolutionize pharmaceutical industry and drug development over the past 5 years. In this expert review, we describe the most frequently used machine learning algorithms in drug development pipelines and the -omics databases well poised to support machine learning and drug discovery. Subsequently, we analyze the emerging new computational approaches to drug discovery and the in silico pipelines for drug repositioning and the synergies among -omics system sciences, AI and machine learning. As with system sciences, AI and machine learning embody a system scale and Big Data driven vision for drug discovery and development. We conclude with a future outlook on the ways in which machine learning approaches can be implemented to buttress and expedite drug discovery and precision medicine. As AI and machine learning are rapidly entering pharmaceutical industry and the art and science of drug development, we need to critically examine the attendant prospects and challenges to benefit patients and public health.

The Need for Multi-Omics Biomarker Signatures in Precision Medicine

Recent advances in omics technologies have led to unprecedented efforts characterizing the molecular changes that underlie the development and progression of a wide array of complex human diseases, including cancer. As a result, multi-omics analyses—which take advantage of these technologies in genomics, transcriptomics, epigenomics, proteomics, metabolomics, and other omics areas—have been proposed and heralded as the key to advancing precision medicine in the clinic. In the field of precision oncology, genomics approaches, and, more recently, other omics analyses have helped reveal several key mechanisms in cancer development, treatment resistance, and recurrence risk, and several of these findings have been implemented in clinical oncology to help guide treatment decisions. However, truly integrated multi-omics analyses have not been applied widely, preventing further advances in precision medicine. Additional efforts are needed to develop the analytical infrastructure necessary to generate, analyze, and annotate multi-omics data effectively to inform precision medicine-based decision-making.

Addressing the Challenges of High-Throughput Cancer Tissue Proteomics for Clinical Application: ProCan

The cancer tissue proteome has enormous potential as a source of novel predictive biomarkers in oncology. Progress in the development of mass spectrometry (MS)-based tissue proteomics now presents an opportunity to exploit this by applying the strategies of comprehensive molecular profiling and big-data analytics that are refined in other fields of 'omics research. ProCan (ProCan is a registered trademark) is a program aiming to generate high-quality tissue proteomic data across a broad spectrum of cancer types. It is based on data-independent acquisition-MS proteomic analysis of annotated tissue samples sourced through collaboration with expert clinical and cancer research groups. The practical requirements of a high-throughput translational research program have shaped the approach that ProCan is taking to address challenges in study design, sample preparation, raw data acquisition, and data analysis. The ultimate goal is to establish a large proteomics knowledge-base that, in combination with other cancer 'omics data, will accelerate cancer research.

MOLI: multi-omics late integration with deep neural networks for drug response prediction

MOTIVATION

Historically, gene expression has been shown to be the most informative data for drug response prediction. Recent evidence suggests that integrating additional omics can improve the prediction accuracy which raises the question of how to integrate the additional omics. Regardless of the integration strategy, clinical utility and translatability are crucial. Thus, we reasoned a multi-omics approach combined with clinical datasets would improve drug response prediction and clinical relevance.

RESULTS

We propose MOLI, a multi-omics late integration method based on deep neural networks. MOLI takes somatic mutation, copy number aberration and gene expression data as input, and integrates them for drug response prediction. MOLI uses type-specific encoding sub-networks to learn features for each omics type, concatenates them into one representation and optimizes this representation via a combined cost function consisting of a triplet loss and a binary cross-entropy loss. The former makes the representations of responder samples more similar to each other and different from the non-responders, and the latter makes this representation predictive of the response values. We validate MOLI on in vitro and in vivo datasets for five chemotherapy agents and two targeted therapeutics. Compared to state-of-the-art single-omics and early integration multi-omics methods, MOLI achieves higher prediction accuracy in external validations. Moreover, a significant improvement in MOLI's performance is observed for targeted drugs when training on a pan-drug input, i.e. using all the drugs with the same target compared to training only on drug-specific inputs. MOLI's high predictive power suggests it may have utility in precision oncology.

AVAILABILITY AND IMPLEMENTATION

https://github.com/hosseinshn/MOLI.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Using proteomics to advance the search for potential biomarkers for preeclampsia: A systematic review and meta-analysis

BACKGROUND

Preeclampsia (PE) is a leading cause of maternal and perinatal morbidity and mortality worldwide. Although predictive multiparametric screening is being developed, it is not applicable to nulliparous women, and is not applied to low-risk women. As PE is considered a heterogenous disorder, it is unlikely that any single multiparametric screening protocol containing a small group of biomarkers could have the required accuracy to predict all PE subgroups. Given the etiology of PE is complex and not fully understood, it begs the question, whether the search for biomarkers based on the predominant view of impaired placentation involving factors predominately implicated in angiogenesis and inflammation, has been too limiting. Here we highlight the enormous potential of state-of-the-art, high-throughput proteomics, to provide a comprehensive and unbiased approach to biomarker identification.

METHODS AND FINDINGS

Our literature search identified 1336 articles; after review, 45 studies with proteomic data from PE women that were eligible for inclusion. From 710 proteins with altered abundance, we identified 13 common circulating proteins, some of which had not been previously considered as prospective biomarkers of PE. An additional search of the literature for original publications testing any of the 13 common proteins using non-proteomic techniques was also undertaken. Strikingly, 9 of these common proteins had been independently evaluated in PE studies as potential biomarkers.

CONCLUSION

This study highlights the potential of using high-throughput data sets, which are comprehensive and without bias, to identify a profile of proteins that may improve predictions of PE and understanding of its etiology. We bring to the attention of the medical and research communities that the strengths and advantages of using data from high-throughput studies for biomarker discovery would be increased dramatically, if first and second trimester samples were collected for proteomics, and if standardized guidelines for patient reporting and data collection were implemented.

Systems Biology Approaches Toward Understanding Primary Mitochondrial Diseases

Primary mitochondrial diseases form one of the most common and severe groups of genetic disease, with a birth prevalence of at least 1 in 5000. These disorders are multi-genic and multi-phenotypic (even within the same gene defect) and span the entire age range from prenatal to late adult onset. Mitochondrial disease typically affects one or multiple high-energy demanding organs, and is frequently fatal in early life. Unfortunately, to date there are no known curative therapies, mostly owing to the rarity and heterogeneity of individual mitochondrial diseases, leading to diagnostic odysseys and difficulties in clinical trial design. This review aims to discuss recent advances and challenges of systems approaches for the study of primary mitochondrial diseases. Although there has been an explosion in the generation of omics data, few studies have progressed toward the integration of multiple levels of omics. It is evident that the integration of different types of data to create a more complete representation of biology remains challenging, perhaps due to the scarcity of available integrative tools and the complexity inherent in their use. In addition, "bottom-up" systems approaches have been adopted for use in the iterative cycle of systems biology: from data generation to model prediction and validation. Primary mitochondrial diseases, owing to their complex nature, will most likely benefit from a multidisciplinary approach encompassing clinical, molecular and computational studies integrated together by systems biology to elucidate underlying pathomechanisms for better diagnostics and therapeutic discovery. Just as next generation sequencing has rapidly increased diagnostic rates from approximately 5% up to 60% over two decades, more recent advancing technologies are encouraging; the generation of multi-omics, the integration of multiple types of data, and the ability to predict perturbations will, ultimately, be translated into improved patient care.

Reproducibility and Crossplatform Validation of Reverse-Phase Protein Array Data

Reverse-phase protein array (RPPA) technology is a high-throughput antibody- and microarray-based approach for the rapid profiling of levels of proteins and protein posttranslational modifications in biological specimens. The technology consumes small amounts of samples, can sensitively detect low-abundance proteins and posttranslational modifications, enables measurements of multiple signaling pathways in parallel, has the capacity to analyze large sample numbers, and offers robust interexperimental reproducibility. These features of RPPA experiments have motivated and enabled the use of RPPA technology in various biomedical, translational, and clinical applications, including the delineation of molecular mechanisms of disease, profiling of druggable signaling pathway activation, and search for new prognostic markers. Owing to the complexity of many of these applications, such as developing multiplex protein assays for diagnostic laboratories or integrating posttranslational modification-level data using large-scale proteogenomic approaches, robust and well-validated data are essential. There are many distinct components of an RPPA workflow, and numerous possible technical setups and analysis parameter options exist. The differences between RPPA platform setups around the world offer opportunities to assess and minimize interplatform variation. Crossplatform validation may also aid in the evaluation of robust, platform-independent protein markers of disease and response to therapy.

Proteomic Profiling of Iron-Treated Ovarian Cells Identifies AKT Activation that Modulates the CLEAR Network

Although iron is essential for cell survival, dysregulated levels can contribute to cancer development or even cell death. The underlying mechanisms mediating these events remain unclear. Herein, proteomic alterations are assessed in iron-treated ovarian cell lines using reverse phase protein array (RPPA) technology and potential functional responses via ingenuity pathway analysis (IPA). Using these approaches, upregulation of pathways modulating organismal death with alterations in mTOR, MAPK, and AKT signaling in HEY ovarian cancer cells in contrast to T80 non-malignant ovarian cells is noted. Since modulation of cell death is mediated in part via microphthalmia-associated transcription factor (MiTF) family, which regulates lysosomal biogenesis and autophagosome formation by upregulating expression of coordinated lysosomal expression and regulation (CLEAR) network, expression changes in these factors in response to iron are investigated. Increased transcription factor EB (TFEB) in T80 (relative to HEY), accompanied by its nuclear translocation and increased CLEAR network gene expression with iron, is identified. Inhibition of AKT alters these responses in contrast to mTOR inhibition, which has little effect. Collectively, these findings support use of RPPA/IPA technology to predict functional responses to iron and further implicate AKT pathway and MiTF members in iron-induced cellular responses in ovarian cells.

From Single Level Analysis to Multi-Omics Integrative Approaches: A Powerful Strategy towards the Precision Oncology

Integration of multi-omics data from different molecular levels with clinical data, as well as epidemiologic risk factors, represents an accurate and promising methodology to understand the complexity of biological systems of human diseases, including cancer. By the extensive use of novel technologic platforms, a large number of multidimensional data can be derived from analysis of health and disease systems. Comprehensive analysis of multi-omics data in an integrated framework, which includes cumulative effects in the context of biological pathways, is therefore eagerly awaited. This strategy could allow the identification of pathway-addiction of cancer cells that may be amenable to therapeutic intervention. However, translation into clinical settings requires an optimized integration of omics data with clinical vision to fully exploit precision cancer medicine. We will discuss the available technical approach and more recent developments in the specific field.

Prediction Methods of Herbal Compounds in Chinese Medicinal Herbs

Chinese herbal medicine has recently gained worldwide attention. The curative mechanism of Chinese herbal medicine is compared with that of western medicine at the molecular level. The treatment mechanism of most Chinese herbal medicines is still not clear. How do we integrate Chinese herbal medicine compounds with modern medicine? Chinese herbal medicine drug-like prediction method is particularly important. A growing number of Chinese herbal source compounds are now widely used as drug-like compound candidates. An important way for pharmaceutical companies to develop drugs is to discover potentially active compounds from related herbs in Chinese herbs. The methods for predicting the drug-like properties of Chinese herbal compounds include the virtual screening method, pharmacophore model method and machine learning method. In this paper, we focus on the prediction methods for the medicinal properties of Chinese herbal medicines. We analyze the advantages and disadvantages of the above three methods, and then introduce the specific steps of the virtual screening method. Finally, we present the prospect of the joint application of various methods.

Machine learning and feature selection for drug response prediction in precision oncology applications

In-depth modeling of the complex interplay among multiple omics data measured from cancer cell lines or patient tumors is providing new opportunities toward identification of tailored therapies for individual cancer patients. Supervised machine learning algorithms are increasingly being applied to the omics profiles as they enable integrative analyses among the high-dimensional data sets, as well as personalized predictions of therapy responses using multi-omics panels of response-predictive biomarkers identified through feature selection and cross-validation. However, technical variability and frequent missingness in input "big data" require the application of dedicated data preprocessing pipelines that often lead to some loss of information and compressed view of the biological signal. We describe here the state-of-the-art machine learning methods for anti-cancer drug response modeling and prediction and give our perspective on further opportunities to make better use of high-dimensional multi-omics profiles along with knowledge about cancer pathways targeted by anti-cancer compounds when predicting their phenotypic responses.

Learning with multiple pairwise kernels for drug bioactivity prediction

Motivation

Many inference problems in bioinformatics, including drug bioactivity prediction, can be formulated as pairwise learning problems, in which one is interested in making predictions for pairs of objects, e.g. drugs and their targets. Kernel-based approaches have emerged as powerful tools for solving problems of that kind, and especially multiple kernel learning (MKL) offers promising benefits as it enables integrating various types of complex biomedical information sources in the form of kernels, along with learning their importance for the prediction task. However, the immense size of pairwise kernel spaces remains a major bottleneck, making the existing MKL algorithms computationally infeasible even for small number of input pairs.

Results

We introduce pairwiseMKL, the first method for time- and memory-efficient learning with multiple pairwise kernels. pairwiseMKL first determines the mixture weights of the input pairwise kernels, and then learns the pairwise prediction function. Both steps are performed efficiently without explicit computation of the massive pairwise matrices, therefore making the method applicable to solving large pairwise learning problems. We demonstrate the performance of pairwiseMKL in two related tasks of quantitative drug bioactivity prediction using up to 167 995 bioactivity measurements and 3120 pairwise kernels: (i) prediction of anticancer efficacy of drug compounds across a large panel of cancer cell lines; and (ii) prediction of target profiles of anticancer compounds across their kinome-wide target spaces. We show that pairwiseMKL provides accurate predictions using sparse solutions in terms of selected kernels, and therefore it automatically identifies also data sources relevant for the prediction problem.

Availability and implementation

Code is available at https://github.com/aalto-ics-kepaco.

Supplementary information

Supplementary data are available at Bioinformatics online.