Differential expression analysis for individual cancer samples based on robust within-sample relative gene expression orderings across multiple profiling platforms

Sample-Wise and Gene-Wise Comparisons Confirm a Greater Similarity of RNA and Protein Expression Data at the Level of Molecular Pathways and Suggest an Approach for the Data Quality Check in High-Throughput Expression Databases

Identification of genes and molecular pathways with congruent profiles in the proteomic and transcriptomic datasets may result in the discovery of promising transcriptomic biomarkers that would be more relevant to phenotypic changes. In this study, we conducted comparative analysis of 943 paired RNA and proteomic profiles obtained for the same samples of seven human cancer types from The Cancer Genome Atlas (TCGA) and NCI Clinical Proteomic Tumor Analysis Consortium (CPTAC) [two major open human cancer proteomic and transcriptomic databases] that included 15,112 protein-coding genes and 1611 molecular pathways. Overall, our findings demonstrated statistically significant improvement of the congruence between RNA and proteomic profiles when performing analysis at the level of molecular pathways rather than at the level of individual gene products. Transition to the molecular pathway level of data analysis increased the correlation to 0.19-0.57 (Pearson) and 0.14-057 (Spearman), or 2-3-fold for some cancer types. Evaluating the gain of the correlation upon transition to the data analysis the pathway level can be used to refine the omics data by identifying outliers that can be excluded from the comparison of RNA and proteomic profiles. We suggest using sample- and gene-wise correlations for individual genes and molecular pathways as a measure of quality of RNA/protein paired molecular data. We also provide a database of human genes, molecular pathways, and samples related to the correlation between RNA and protein products to facilitate an exploration of new cancer transcriptomic biomarkers and molecular mechanisms at different levels of human gene expression.

A pyroptosis-related lncRNA-based prognostic index for hepatocellular carcinoma by relative expression orderings

Background

Hepatocellular carcinoma (HCC) is an invasive malignant tumor, and pyroptosis makes an important contribution to the pathology and progression of liver cancer. Many prognostic models have been proposed for HCC based on the quantitative expression level of candidate genes, which are unsuitable for clinical application due to their vulnerability against experimental batch effects. The aim of this study was to develop a novel pyroptosis-related long non-coding RNA (lncRNA)-based prognostic index (PLPI) for HCC based on relative expression orderings (REOs).

Methods

Firstly, the pyroptosis-related lncRNAs were identified through the Wilcoxon rank-sum test and gene co-expression analyses. Then, the novel prognostic model PLPI was constructed by pyroptosis-related lncRNA pairs, which were identified by multiple machine learning algorithms. Gene set enrichment, somatic mutation, and drug sensitivity analyses were conducted to measure the differences between high- and low-risk patients. Multiple immune analyses were used to explore the association between PLPI and the immunological microenvironment.

Results

In this study, a novel prognostic model PLPI based on 10 pyroptosis-related lncRNA pairs was constructed, which was proven to be an independent prognostic risk factor. The receiver operating characteristic (ROC) curves showed that the model had a good prognostic ability in the training, testing, and external set, respectively [5-year area under the curve (AUC) =0.73, 5-year AUC =0.81, 4-year AUC =0.79]. The results of survival, somatic mutation, and immune analyses showed that the patients in the low-risk group had a better prognosis, lower rates of somatic mutation, and better immune cell infiltration. Personalized chemotherapeutic drugs were also identified for the patients with HCC.

Conclusions

The novel PLPI not only greatly predicted the prognosis of patients with HCC but could also offer novel ideas and approaches for the therapeutic management of HCC.

Excavation of gene markers associated with pancreatic ductal adenocarcinoma based on interrelationships of gene expression

Pancreatic ductal adenocarcinoma (PDAC) accounts for 95% of all pancreatic cancer cases, posing grave challenges to its diagnosis and treatment. Timely diagnosis is pivotal for improving patient survival, necessitating the discovery of precise biomarkers. An innovative approach was introduced to identify gene markers for precision PDAC detection. The core idea of our method is to discover gene pairs that display consistent opposite relative expression and differential co-expression patterns between PDAC and normal samples. Reversal gene pair analysis and differential partial correlation analysis were performed to determine reversal differential partial correlation (RDC) gene pairs. Using incremental feature selection, the authors refined the selected gene set and constructed a machine-learning model for PDAC recognition. As a result, the approach identified 10 RDC gene pairs. And the model could achieve a remarkable accuracy of 96.1% during cross-validation, surpassing gene expression-based models. The experiment on independent validation data confirmed the model's performance. Enrichment analysis revealed the involvement of these genes in essential biological processes and shed light on their potential roles in PDAC pathogenesis. Overall, the findings highlight the potential of these 10 RDC gene pairs as effective diagnostic markers for early PDAC detection, bringing hope for improving patient prognosis and survival.

Identification of cancer risk assessment signature in patients with chronic obstructive pulmonary disease and exploration of the potential key genes

It is essential to assess the cancer risk for patients with chronic obstructive pulmonary disease (COPD). Comparing gene expression data from patients with lung cancer (a total of 506 samples) and those with cancer-adjacent normal lung tissues (a total of 370 samples), we generated a qualitative transcriptional signature consisting of 2046 gene pairs. The signature was verified in an evaluation dataset comprising 18 subjects with severe disease and 52 subjects with moderate disease (Wilcoxon rank-sum test; p = 7.33 × 10^-5). Similar results were obtained in other independent datasets. Among the gene pairs in the signature, 326 COPD stage-related gene pairs were identified based on Spearman's rank correlation tests and those gene pairs comprised 368 unique genes. Of these 368 genes, 16 genes were significantly dysregulated in COPD rat model data compared with control data. Some of these genes (Dhx16, Upf2, Notch3, Sec61a1, Dyrk2, and Hmmr) were altered when the COPD rat model was treated with traditional Chinese medicines (TCM), including Bufei Yishen formula, Bufei Jianpi formula, and Yiqi Zishen formula. Overall, the signature could predict the cancer incidence-risk of COPD and the identified key genes might provide guidance regarding both the treatment of COPD using TCM and the prevention of cancer in patients with COPD. KEY MESSAGESA cancer risk assessment signature was identified in patients with COPD.The signature is insensitive to batch effects and is well verified.COPD key genes identified in this study might play a crucial role in TCM treatment and cancer prevention.

Metabolism-Related Gene Pairs to Predict the Clinical Outcome and Molecular Characteristics of Early Hepatocellular Carcinoma

Simple Summary

After surgery, about 60–70% of early hepatocellular carcinoma patients suffer from relapse within 5 years, hindering long-term survival. Clinical and pathologic features cannot provide an accurate evaluation. We aimed to construct a stratification model from the metabolic aspect to predict the clinical outcome and reveal the molecular characteristics of different prognostic subgroups. An individualized metabolic signature of 10 gene pairs was developed from 250 early HCCs and validated in 311 samples from different datasets. The signature stratified early HCC cases one-by-one into two risk groups with different survival rates. The molecular characteristics of the two risk groups were analyzed by multi-omics data. The relationships with proliferation, immunity, and drug benefits were summarized. The signature was further validated in 47 institutional transcriptomic HCC samples and 101 public proteomic samples.

Abstract

Recurrence is the main factor affecting the prognosis of early hepatocellular carcinoma (HCC), which is not accurately evaluated by clinical indicators. The metabolic heterogeneity of HCC hints at the possibility of constructing a stratification model to predict the clinical outcome. On the basis of the relative expression orderings of 2939 metabolism-related genes, an individualized signature with 10 metabolism-related gene pairs (10-GPS) was developed from 250 early HCC samples in the discovery datasets, which stratified HCC patients into the high- and low-risk subgroups with significantly different survival rates. The 10-GPS was validated in 311 public transcriptomic samples from two independent validation datasets. A nomogram that included the 10-GPS, age, gender, and stage was constructed for eventual clinical evaluation. The low-risk group was characterized by lower proliferation, higher metabolism, increased activated immune microenvironment, and lower TIDE scores, suggesting a better response to immunotherapy. The high-risk group displayed hypomethylation, higher copy number alterations, mutations, and more overexpression of immune-checkpoint genes, which might jointly lead to poor outcomes. The prognostic accuracy of the 10-GPS was further validated in 47 institutional transcriptomic samples and 101 public proteomic samples. In conclusion, the 10-GPS is a robust predictor of the clinical outcome for early HCC patients and could help evaluate prognosis and characterize molecular heterogeneity.

Prognostic value of a microRNA-pair signature in laryngeal squamous cell carcinoma patients

PURPOSE

Predicting the prognosis in laryngeal squamous cell carcinoma (LSCC) patients will improve clinical decision-making. Here, we aimed to identify a qualitative signature based on the within-sample relative expression orderings (REOs) of microRNA (miRNA) pairs to predict the overall survival (OS) of LSCC patients.

METHODS

First, we constructed non-repeating miRNA pairs based on differentially expressed miRNAs (DEmiRNAs) between LSCC and normal tissues. Then, we applied a bootstrap-based feature selection method to identify a robust miRNA-pair signature. The bootstrap-based feature selection improved the stability of feature selection by an ensemble based on the data perturbation. Furthermore, a series of bioinformatics analyses were carried out to explore the potential mechanisms of the signature and potential drug targets for LSCC.

RESULTS

Based on the REOs of miRNA pairs, we identified a qualitative signature that consisted of 12 miRNA pairs. The constructed signature has good performance in predicting the OS of LSCC patients. It is robust against batch effects and more suitable for individual clinical applications. Furthermore, we identified several hub genes that may be potential drug targets for LSCC.

CONCLUSION

Overall, our findings provided a promising signature for predicting the OS of LSCC patients.

Transcriptomic Portraits and Molecular Pathway Activation Features of Adult Spinal Intramedullary Astrocytomas

In this study, we report 31 spinal intramedullary astrocytoma (SIA) RNA sequencing (RNA-seq) profiles for 25 adult patients with documented clinical annotations. To our knowledge, this is the first clinically annotated RNA-seq dataset of spinal astrocytomas derived from the intradural intramedullary compartment. We compared these tumor profiles with the previous healthy central nervous system (CNS) RNA-seq data for spinal cord and brain and identified SIA-specific gene sets and molecular pathways. Our findings suggest a trend for SIA-upregulated pathways governing interactions with the immune cells and downregulated pathways for the neuronal functioning in the context of normal CNS activity. In two patient tumor biosamples, we identified diagnostic KIAA1549-BRAF fusion oncogenes, and we also found 16 new SIA-associated fusion transcripts. In addition, we bioinformatically simulated activities of targeted cancer drugs in SIA samples and predicted that several tyrosine kinase inhibitory drugs and thalidomide analogs could be potentially effective as second-line treatment agents to aid in the prevention of SIA recurrence and progression.

The prognostic and clinical significance of IFI44L aberrant downregulation in patients with oral squamous cell carcinoma

Background

It is a basic task in high-throughput gene expression profiling studies to identify differentially expressed genes (DEGs) between two phenotypes. RankComp, an algorithm, could analyze the highly stable within-sample relative expression orderings (REOs) of gene pairs in a particular type of human normal tissue that are widely reversed in the cancer condition, thereby detecting DEGs for individual disease samples measured by a particular platform.

Methods

In the present study, Gene Expression Omnibus (GEO) Series (GSE) GSE75540, GSE138206 were downloaded from GEO, by analyzing DEGs in oral squamous cell carcinoma based on online datasets using the RankComp algorithm, using the Kaplan-Meier survival analysis and Cox regression analysis to survival analysis, Gene Set Enrichment Analysis (GSEA) to explore the potential molecular mechanisms underlying.

Results

We identified 6 reverse gene pairs with stable REOs. All the 12 genes in these 6 reverse gene pairs have been reported to be associated with cancers. Notably, lower Interferon Induced Protein 44 Like (IFI44L) expression was associated with poorer overall survival (OS) and Disease-free survival (DFS) in oral squamous cell carcinoma patients, and IFI44L expression showed satisfactory predictive efficiency by receiver operating characteristic (ROC) curve. Moreover, low IFI44L expression was identified as risk factors for oral squamous cell carcinoma patients’ OS. IFI44L downregulation would lead to the activation of the FRS-mediated FGFR1, FGFR3, and downstream signaling pathways, and might play a role in the PI3K-FGFR cascades.

Conclusions

Collectively, we identified 6 reverse gene pairs with stable REOs in oral squamous cell carcinoma, which might serve as gene signatures playing a role in the diagnosis in oral squamous cell carcinoma. Moreover, high expression of IFI44L, one of the DEGs in the 6 reverse gene pairs, might be associated with favorable prognosis in oral squamous cell carcinoma patients and serve as a tumor suppressor by acting on the FRS-mediated FGFR signaling.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-021-09058-y.

A transcriptional signature detects homologous recombination deficiency in pancreatic cancer at the individual level

Pancreatic cancer (PC) with homologous recombination deficiency (HRD) has been reported to benefit from poly ADP-ribose polymerase (PARP) inhibitors. However, accurate identification of HRD status for PC patients from the transcriptional level is still a great challenge. Here, based on a relative expression ordering (REO)-based algorithm, we developed an HRD signature including 24 gene pairs (24-GPS) using PC transcriptional profiles from The Cancer Genome Atlas (TCGA). HRD samples classified by 24-GPS showed worse overall survival (p = 4.4E-3 for TCGA; p = 1.2E-3 for International Cancer Genome Consortium-Australia cohort; p = 6.4E-2 for GSE17891; p = 7.5E-2 for GSE57495) and higher HRD scores than non-HRD samples (p = 1.4E-4). HRD samples showed highly unstable genomic characteristics and also displayed HRD-related alterations at the epigenomic and proteomic levels. Moreover, HRD cell lines identified by 24-GPS tended to be sensitive to PARP inhibitors (p = 6.6E-2 for olaparib; p = 2.6E-3 for niraparib). Compared with the non-HRD group, the HRD group presented lower immune scores and CD4/CD8 T cell infiltration proportion. Interestingly, PC tumor cells with co-inhibition of PARP-related genes and ATR showed reduced survival ability. In conclusion, 24-GPS can robustly identify PC patients with HRD status at the individualized level.

Large-Scale Transcriptomics-Driven Approach Revealed Overexpression of CRNDE as a Poor Survival Prognosis Biomarker in Glioblastoma

Glioblastoma is the most common and malignant brain malignancy worldwide, with a 10-year survival of only 0.7%. Aggressive multimodal treatment is not enough to increase life expectancy and provide good quality of life for glioblastoma patients. In addition, despite decades of research, there are no established biomarkers for early disease diagnosis and monitoring of patient response to treatment. High throughput sequencing technologies allow for the identification of unique molecules from large clinically annotated datasets. Thus, the aim of our study was to identify significant molecular changes between short- and long-term glioblastoma survivors by transcriptome RNA sequencing profiling, followed by differential pathway-activation-level analysis. We used data from the publicly available repositories The Cancer Genome Atlas (TCGA; number of annotated cases = 135) and Chinese Glioma Genome Atlas (CGGA; number of annotated cases = 218), and experimental clinically annotated glioblastoma tissue samples from the Institute of Pathology, Faculty of Medicine in Ljubljana corresponding to 2-58 months overall survival (n = 16). We found one differential gene for long noncoding RNA CRNDE whose overexpression showed correlation to poor patient OS. Moreover, we identified overlapping sets of congruently regulated differential genes involved in cell growth, division, and migration, structure and dynamics of extracellular matrix, DNA methylation, and regulation through noncoding RNAs. Gene ontology analysis can provide additional information about the function of protein- and nonprotein-coding genes of interest and the processes in which they are involved. In the future, this can shape the design of more targeted therapeutic approaches.

Long non-coding RNA pairs to assist in diagnosing sepsis

Background

Sepsis is the major cause of death in Intensive Care Unit (ICU) globally. Molecular detection enables rapid diagnosis that allows early intervention to minimize the death rate. Recent studies showed that long non-coding RNAs (lncRNAs) regulate proinflammatory genes and are related to the dysfunction of organs in sepsis. Identifying lncRNA signature with absolute abundance is challenging because of the technical variation and the systematic experimental bias.

Results

Cohorts (n = 768) containing whole blood lncRNA profiling of sepsis patients in the Gene Expression Omnibus (GEO) database were included. We proposed a novel diagnostic strategy that made use of the relative expressions of lncRNA pairs, which are reversed between sepsis patients and normal controls (eg. lncRNA_i > lncRNA_j in sepsis patients and lncRNA_i < lncRNA_j in normal controls), to identify 14 lncRNA pairs as a sepsis diagnostic signature. The signature was then applied to independent cohorts (n = 644) to evaluate its predictive performance across different ages and normalization methods. Comparing to common machine learning models and existing signatures, SepSigLnc consistently attains better performance on the validation cohorts from the same age group (AUC = 0.990 & 0.995 in two cohorts) and across different groups (AUC = 0.878 on average), as well as cohorts processed by an alternative normalization method (AUC = 0.953 on average). Functional analysis demonstrates that the lncRNA pairs in SepsigLnc are functionally similar and tend to implicate in the same biological processes including cell fate commitment and cellular response to steroid hormone stimulus.

Conclusion

Our study identified 14 lncRNA pairs as signature that can facilitate the diagnosis of septic patients at an intervenable point when clinical manifestations are not dramatic. Also, the computational procedure can be generalized to a standard procedure for discovering diagnostic molecule signatures.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-021-07576-4.

Identification of Genes Universally Differentially Expressed in Gastric Cancer

Owing to the remarkable heterogeneity of gastric cancer (GC), population-level differentially expressed genes (DEGs) identified using case-control comparison cannot indicate the dysregulated frequency of each DEG in GC. In this work, first, the individual-level DEGs were identified for 1,090 GC tissues without paired normal tissues using the RankComp method. Second, we directly compared the gene expression in a cancer tissue to that in paired normal tissue to identify individual-level DEGs among 448 paired cancer-normal gastric tissues. We found 25 DEGs to be dysregulated in more than 90% of 1,090 GC tissues and also in more than 90% of 448 GC tissues with paired normal tissues. The 25 genes were defined as universal DEGs for GC. Then, we measured 24 paired cancer-normal gastric tissues by RNA-seq to validate them further. Among the universal DEGs, 4 upregulated genes (BGN, E2F3, PLAU, and SPP1) and 1 downregulated gene (UBL3) were found to be cancer genes already documented in the COSMIC or F-Census databases. By analyzing protein-protein interaction networks, we found 12 universally upregulated genes, and we found that their 284 direct neighbor genes were significantly enriched with cancer genes and key biological pathways related to cancer, such as the MAPK signaling pathway, cell cycle, and focal adhesion. The 13 universally downregulated genes and 16 direct neighbor genes were also significantly enriched with cancer genes and pathways related to gastric acid secretion. These universal DEGs may be of special importance to GC diagnosis and treatment targets, and they may make it easier to study the molecular mechanisms underlying GC.

A Qualitative Transcriptional Signature for the Risk Assessment of Precancerous Colorectal Lesions

It is meaningful to assess the risk of cancer incidence among patients with precancerous colorectal lesions. Comparing the within-sample relative expression orderings (REOs) of colorectal cancer patients measured by multiple platforms with that of normal colorectal tissues, a qualitative transcriptional signature consisting of 1,840 gene pairs was identified in the training data. Within an evaluation dataset of 16 active and 18 inactive (remissive) ulcerative colitis subjects, the median incidence risk score of colorectal carcinoma was 0.6402 in active ulcerative colitis subjects, significantly higher than that in remissive subjects (0.3114). Evaluation of two other independent datasets yielded similar results. Moreover, we found that the score significantly positively correlated with the degree of dysplasia in the case of colorectal adenomas. In the merged dataset, the median incidence risk score was 0.9027 among high-grade adenoma samples, significantly higher than that among low-grade adenomas (0.8565). In summary, the developed incidence risk score could well predict the incidence risk of precancerous colorectal lesions and has value in clinical application.

Transcriptomic analysis identifies organ-specific metastasis genes and pathways across different primary sites

Background

Metastasis is the most devastating stage of cancer progression and often shows a preference for specific organs.

Methods

To reveal the mechanisms underlying organ-specific metastasis, we systematically analyzed gene expression profiles for three common metastasis sites across all available primary origins. A rank-based method was used to detect differentially expressed genes between metastatic tumor tissues and corresponding control tissues. For each metastasis site, the common differentially expressed genes across all primary origins were identified as organ-specific metastasis genes.

Results

Pathways enriched by these genes reveal an interplay between the molecular characteristics of the cancer cells and those of the target organ. Specifically, the neuroactive ligand-receptor interaction pathway and HIF-1 signaling pathway were found to have prominent roles in adapting to the target organ environment in brain and liver metastases, respectively. Finally, the identified organ-specific metastasis genes and pathways were validated using a primary breast tumor dataset. Survival and cluster analysis showed that organ-specific metastasis genes and pathways tended to be expressed uniquely by a subgroup of patients having metastasis to the target organ, and were associated with the clinical outcome.

Conclusions

Elucidating the genes and pathways underlying organ-specific metastasis may help to identify drug targets and develop treatment strategies to benefit patients.

A Five-Gene-Pair-Based Prognostic Signature for Predicting the Relapse Risk of Early Stage ER+ Breast Cancer

About 20-30% of early-stage breast cancer patients suffer relapses after surgery. To identify such high-risk patients, many signatures have been reported, but they lack robustness in data measured on different platforms. Here, we developed a signature which is robust across multiple profiling platforms, and identified reproducible omics features characterizing metastasis of estrogen receptor (ER)-positive breast cancer from the Gene Expression Omnibus database with the aid of the signature. Based on the stable within-sample relative expression orderings (REOs), we constructed a signature consisting of five gene pairs, named 5-GPS, whose REOs were significantly correlated with relapse-free survival using the univariate Cox regression model. Using 5-GPS, patients were classified into the low-risk and high-risk groups. Patients in the high-risk group have worse survival compared to those in the low-risk group using Kaplan-Meier curve analysis with the log-rank test. Applying 5-GPS to the RNA-sequencing data of stage I-IV breast cancer samples archived in The Cancer Genome Atlas (TCGA), we found that the proportion of the high-risk patients increases with the stage. The proposed REO-based signature shows potential in identifying early-stage ER+ breast cancer patients with high risk of relapse after surgery.

Early Diagnosis of Pancreatic Ductal Adenocarcinoma by Combining Relative Expression Orderings With Machine-Learning Method

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive and lethal cancer deeply affecting human health. Diagnosing early-stage PDAC is the key point to PDAC patients' survival. However, the biomarkers for diagnosing early PDAC are inexact in most cases. Therefore, it is highly desirable to identify an effective PDAC diagnostic biomarker. In the current work, we designed a novel computational approach based on within-sample relative expression orderings (REOs). A feature selection technique called minimum redundancy maximum relevance was used to pick out optimal REOs. We then compared the performances of different classification algorithms for discriminating PDAC and its adjacent normal tissues from non-PDAC tissues. The support vector machine algorithm is the best one for identifying early PDAC diagnostic biomarker. At first, a signature composed of nine gene pairs was acquired from microarray gene expression data sets. These gene pairs could produce satisfactory classification accuracy up to 97.53% in fivefold cross-validation. Subsequently, two types of data from diverse platforms, namely, microarray and RNA-Seq, were used to validate this signature. For microarray data, all (100.00%) of 115 PDAC tissues and all (100.00%) of 31 PDAC adjacent normal tissues were correctly recognized as PDAC. In addition, 88.24% of 17 non-PDAC (normal or pancreatitis) tissues were correctly classified. For the RNA-Seq data, all (100.00%) of 177 PDAC tissues and all (100.00%) of 4 PDAC adjacent normal tissues were correctly recognized as PDAC. Validation results demonstrated that the signature had a good cross-platform effect for early detection of PDAC. This work developed a new robust signature that might be a promising biomarker for early PDAC diagnosis.

PenDA, a rank-based method for personalized differential analysis: Application to lung cancer

The hopes of precision medicine rely on our capacity to measure various high-throughput genomic information of a patient and to integrate them for personalized diagnosis and adapted treatment. Reaching these ambitious objectives will require the development of efficient tools for the detection of molecular defects at the individual level. Here, we propose a novel method, PenDA, to perform Personalized Differential Analysis at the scale of a single sample. PenDA is based on the local ordering of gene expressions within individual cases and infers the deregulation status of genes in a sample of interest compared to a reference dataset. Based on realistic simulations of RNA-seq data of tumors, we showed that PenDA outcompetes existing approaches with very high specificity and sensitivity and is robust to normalization effects. Applying the method to lung cancer cohorts, we observed that deregulated genes in tumors exhibit a cancer-type-specific commitment towards up- or down-regulation. Based on the individual information of deregulation given by PenDA, we were able to define two new molecular histologies for lung adenocarcinoma cancers strongly correlated to survival. In particular, we identified 37 biomarkers whose up-regulation lead to bad prognosis and that we validated on two independent cohorts. PenDA provides a robust, generic tool to extract personalized deregulation patterns that can then be used for the discovery of therapeutic targets and for personalized diagnosis. An open-access, user-friendly R package is available at https://github.com/bcm-uga/penda.

The hopes of precision medicine rely on our capacity to measure individual molecular information for personalized diagnosis and treatment. These challenging perspectives will be only possible with the development of efficient methodological tools to identify patient-specific molecular defects from the many precise molecular information that one can access at the single-individual, single tissue or even single-cell levels. Such methods will provide a better understanding of disease-specific biological mechanisms and will promote the development of personalized therapeutic strategies. Here we describe a novel method, named PenDA, to perform differential analysis of gene expression at the individual level. Based on a realistic benchmark of simulated tumors, we demonstrated that PenDA reaches very high efficiency in detecting sample-specific deregulated genes. We then applied the method to two large cohorts associated with lung cancer. A detailed statistical analysis of the results allowed to isolate genes with specific deregulation patterns, like genes that are up-regulated in all tumors or genes that are expressed but never deregulated in any tumors. Given their specificities, these genes are likely to be of interest in therapeutic research. In particular, we were able to identified 37 new biomarkers associated to bad prognosis that we validated on two independent cohorts.

A qualitative transcriptional prognostic signature for patients with stage I-II pancreatic ductal adenocarcinoma

Accurately prognostic evaluation of patients with stage I-II pancreatic ductal adenocarcinoma (PDAC) is of importance to treatment decision and patient management. Most previously reported prognostic signatures were based on risk scores summarized from quantitative expression measurements of signature genes, which are susceptible to experimental batch effects and impractical for clinical applications. Based on the within-sample relative expression orderings of genes, we developed a robust qualitative transcriptional prognostic signature, consisting of 64 gene pairs (64-GPS), to predict the overall survival (OS) of 161 stage I-II PDAC patients in the training dataset who were treated with surgery only. Samples were classified into the high-risk group when at least 25 of 64 gene pairs suggested it was at high risk. The signature was successfully validated in 324 samples from 6 independent datasets produced by different laboratories. All samples in the low-risk group had significantly better OS than samples in the high-risk group. Multivariate Cox regression analyses showed that the 64-GPS remained significantly associated with the OS of patients after adjusting available clinical factors. Transcriptomic analysis of the 2 prognostic subgroups showed that the differential expression signals were highly reproducible in all datasets, whereas the differences between samples grouped by the TNM staging system were weak and irreproducible. The epigenomic analysis showed that the epigenetic alternations may cause consistently transcriptional changes between the 2 different prognostic groups. The genomic analysis revealed that mutation‑induced disturbances in several key genes, such as LRMDA, MAPK10, and CREBBP, might lead to poor prognosis for PDAC patients. Conclusively, the 64-GPS can robustly predict the prognosis of patients with stage I-II PDAC, which provides theoretical basis for clinical individualized treatment.

A qualitative transcriptional signature to reclassify histological grade of ER-positive breast cancer patients

Background

Histological grade (HG) is commonly adopted as a prognostic factor for ER-positive breast cancer patients. However, HG evaluation methods, such as the pathological Nottingham grading system, are highly subjective with only 50–85% inter-observer agreements. Specifically, the subjectivity in the pathological assignment of the intermediate grade (HG2) breast cancers, comprising of about half of breast cancer cases, results in uncertain disease outcomes prediction. Here, we developed a qualitative transcriptional signature, based on within-sample relative expression orderings (REOs) of gene pairs, to define HG1 and HG3 and reclassify pathologically-determined HG2 (denoted as pHG2) breast cancer patients.

Results

From the gene pairs with significantly stable REOs in pathologically-determined HG1 (denoted as pHG1) samples and reversely stable REOs in pathologically-determined HG3 (denoted as pHG3) samples, concordantly identified from seven datasets, we extracted a signature which could determine the HG state of samples through evaluating whether the within-sample REOs match with the patterns of the pHG1 REOs or pHG3 REOs. A sample was classified into the HG3 group if at least a half of the REOs of the 10 gene pairs signature within this sample voted for HG3; otherwise, HG1. Using four datasets including samples of early stage (I–II) ER-positive breast cancer patients who accepted surgery only, we validated that this signature was able to reclassify pHG2 patients into HG1 and HG3 groups with significantly different survival time. For the original pHG1 and pHG3 patients, the signature could also more accurately and objectively stratify them into distinct prognostic groups. And the up-regulated and down down-regulated genes in HG1 compared with HG3 involved in cell proliferation and extracellular signal transduction pathways respectively. By comparing with existing signatures, 10-GPS was with prognostic significance and was more aligned with survival of patients especially for pHG2 samples.

Conclusions

The transcriptional qualitative signature can provide an objective assessment of HG states of ER-positive breast cancer patients, especially for reclassifying patients with pHG2, to assist decision making on clinical therapy.

Early Diagnosis of Hepatocellular Carcinoma Using Machine Learning Method

Hepatocellular carcinoma (HCC) is a serious cancer which ranked the fourth in cancer-related death worldwide. Hence, more accurate diagnostic models are urgently needed to aid the early HCC diagnosis under clinical scenarios and thus improve HCC treatment and survival. Several conventional methods have been used for discriminating HCC from cirrhosis tissues in patients without HCC (CwoHCC). However, the recognition successful rates are still far from satisfactory. In this study, we applied a computational approach that based on machine learning method to a set of microarray data generated from 1091 HCC samples and 242 CwoHCC samples. The within-sample relative expression orderings (REOs) method was used to extract numerical descriptors from gene expression profiles datasets. After removing the unrelated features by using maximum redundancy minimum relevance (mRMR) with incremental feature selection, we achieved “11-gene-pair” which could produce outstanding results. We further investigated the discriminate capability of the “11-gene-pair” for HCC recognition on several independent datasets. The wonderful results were obtained, demonstrating that the selected gene pairs can be signature for HCC. The proposed computational model can discriminate HCC and adjacent non-cancerous tissues from CwoHCC even for minimum biopsy specimens and inaccurately sampled specimens, which can be practical and effective for aiding the early HCC diagnosis at individual level.

An absolute human stemness index associated with oncogenic dedifferentiation

The progression of cancer is accompanied by the acquisition of stemness features. Many stemness evaluation methods based on transcriptional profiles have been presented to reveal the relationship between stemness and cancer. However, instead of absolute stemness index values-the values with certain range-these methods gave the values without range, which makes them unable to intuitively evaluate the stemness. Besides, these indices were based on the absolute expression values of genes, which were found to be seriously influenced by batch effects and the composition of samples in the dataset. Recently, we have showed that the signatures based on the relative expression orderings (REOs) of gene pairs within a sample were highly robust against these factors, which makes that the REO-based signatures have been stably applied in the evaluations of the continuous scores with certain range. Here, we provided an absolute REO-based stemness index to evaluate the stemness. We found that this stemness index had higher correlation with the culture time of the differentiated stem cells than the previous stemness index. When applied to the cancer and normal tissue samples, the stemness index showed its significant difference between cancers and normal tissues and its ability to reveal the intratumor heterogeneity at stemness level. Importantly, higher stemness index was associated with poorer prognosis and greater oncogenic dedifferentiation reflected by histological grade. All results showed the capability of the REO-based stemness index to assist the assignment of tumor grade and its potential therapeutic and diagnostic implications.

Qualitative transcriptional signature for predicting pathological response of colorectal cancer to FOLFOX therapy

FOLFOX (5‐fluorouracil, leucovorin and oxaliplatin) is one of the main chemotherapy regimens for colorectal cancer (CRC), but only half of CRC patients respond to this regimen. Using gene expression profiles of 96 metastatic CRC patients treated with FOLFOX, we first selected gene pairs whose within‐sample relative expression orderings (REO) were significantly associated with the response to FOLFOX using the exact binomial test. Then, from these gene pairs, we applied an optimization procedure to obtain a subset that achieved the largest F‐score in predicting pathological response of CRC to FOLFOX. The REO‐based qualitative transcriptional signature, consisting of five gene pairs, was developed in the training dataset consisting of 96 samples with an F‐score of 0.90. In an independent test dataset consisting of 25 samples with the response information, an F‐score of 0.82 was obtained. In three other independent survival datasets, the predicted responders showed significantly better progression‐free survival than the predicted non‐responders. In addition, the signature showed a better predictive performance than two published FOLFOX signatures across different datasets and is more suitable for CRC patients treated with FOLFOX than 5‐fluorouracil‐based signatures. In conclusion, the REO‐based qualitative transcriptional signature can accurately identify metastatic CRC patients who may benefit from the FOLFOX regimen.

A qualitative transcriptional signature for predicting microsatellite instability status of right-sided Colon Cancer

Background

Microsatellite instability (MSI) accounts for about 15% of colorectal cancer and is associated with prognosis. Today, MSI is usually detected by polymerase chain reaction amplification of specific microsatellite markers. However, the instability is identified by comparing the length of microsatellite repeats in tumor and normal samples. In this work, we developed a qualitative transcriptional signature to individually predict MSI status for right-sided colon cancer (RCC) based on tumor samples.

Results

Using RCC samples, based on the relative expression orderings (REOs) of gene pairs, we extracted a signature consisting of 10 gene pairs (10-GPS) to predict MSI status for RCC through a feature selection process. A sample is predicted as MSI when the gene expression orderings of at least 7 gene pairs vote for MSI; otherwise the microsatellite stability (MSS). The classification performance reached the largest F-score in the training dataset. This signature was verified in four independent datasets of RCCs with the F-scores of 1, 0.9630, 0.9412 and 0.8798, respectively. Additionally, the hierarchical clustering analyses and molecular features also supported the correctness of the reclassifications of the MSI status by 10-GPS.

Conclusions

The qualitative transcriptional signature can be used to classify MSI status of RCC samples at the individualized level.

A qualitative transcriptional signature for the early diagnosis of colorectal cancer

Currently, using biopsy specimens for the early diagnosis of colorectal cancer (CRC) is not entirely reliable due to insufficient sampling amount and inaccurate sampling location. Thus, it is necessary to develop a signature that can accurately identify patients with CRC under these clinical scenarios. Based on the relative expression orderings of genes within individual samples, we developed a qualitative transcriptional signature to discriminate CRC tissues, including CRC adjacent normal tissues from non-CRC individuals. The signature was validated using multiple microarray and RNA sequencing data from different sources. In the training data, a signature consisting of 7 gene pairs was identified. It was well validated in both biopsy and surgical resection specimens from multiple datasets measured by different platforms. For biopsy specimens, 97.6% of 42 CRC tissues and 94.5% of 163 non-CRC (normal or inflammatory bowel disease) tissues were correctly classified. For surgically resected specimens, 99.5% of 854 CRC tissues and 96.3% of 81 CRC adjacent normal tissues were correctly identified as CRC. Notably, we additionally measured 33 CRC biopsy specimens by the Affymetrix platform and 13 CRC surgical resection specimens, with different proportions of tumor epithelial cells, ranging from 40% to 100%, by the RNA sequencing platform, and all these samples were correctly identified as CRC. The signature can be used for the early diagnosis of CRC, which is also suitable for minimum biopsy specimens and inaccurately sampled specimens, and thus has potential value for clinical application.

Identification of genes with universally upregulated or downregulated expressions in colorectal cancer

BACKGROUND AND AIM

Differentially expressed (DE) genes detected at the population-level through case-control comparison provide no information on the dysregulation frequencies of DE genes in a cancer. In this work, we aimed to identify the genes with universally upregulated or downregulated expressions in colorectal cancer (CRC).

METHODS

We firstly clarified that DE genes in an individual cancer tissue should be the disease-relevant genes, which are dysregulated in the cancer tissue in comparison with its own previously normal state. Then, we identified DE genes at the individual level for 2233 CRC samples collected from multiple data sources using the RankComp algorithm.

RESULTS

We found 26 genes that were upregulated or downregulated in almost all the 2233 CRC samples and validated the results using 124 CRC tissues with paired adjacent normal tissues. Especially, we found that two genes (AJUBA and EGFL6), upregulated universally in CRC tissues, were extremely lowly expressed in normal colorectal tissues, which were considered to be oncogenes in CRC oncogenesis and development. Oppositely, we found that one gene (LPAR1), downregulated universally in CRC tissues, was silenced in CRC tissues but highly expressed in normal colorectal tissues, which were considered to be tumor suppressor genes in CRC. Functional evidences revealed that these three genes may induce CRC through deregulating pathways for ribosome biogenesis, cell proliferation, and cell cycle.

CONCLUSIONS

In conclusion, the individual-level DE genes analysis can help us find genes dysregulated universally in CRC tissues, which may be important diagnostic biomarkers and therapy targets.

Qualitative transcriptional signatures for evaluating the maturity degree of pluripotent stem cell-derived cardiomyocytes

Background

Pluripotent stem cell-derived cardiomyocytes (PSC-CMs) are widely used models for regenerative medicine and disease research. However, PSC-CMs are usually immature in morphology and functionality and the maturity of PSC-CMs could not be determined accurately. In order to reasonably interpret the experimental results obtained by PSC-CMs, it is necessary to evaluate the maturity of PSC-CMs and find the key genes related to maturation.

Methods

Using the gene expression profiles of normal adult cardiac tissue and embryonic stem cell (ESC) samples, we identified gene pairs with identically relative expression orderings (REOs) within adult cardiac tissue but reversely identical in ESCs. Then, for a PSC-CM model, we calculated the maturity score as the percentage of these gene pairs that exhibit the same REOs in adult cardiac tissue. Lastly, the CellComp method was used to identify the maturation-related genes.

Results

The maturity score increased gradually from 0.8401 for 18-week fetal cardiac tissue to 0.9997 for adult cardiac tissue. For four human PSC-CM models, the mature scores increased with prolonged culture time but were all below 0.8. The genes involved in energy metabolism, angiogenesis, immunity, and proliferation were dysregulated in the 1-year PSC-CMs compared with adult cardiac tissue.

Conclusion

We proposed a qualitative transcriptional signature to score the maturity degree of PSC-CMs. This score can reasonably track the maturity of PSC-CMs and be used to compare different PSC-CM culture methods.

Electronic supplementary material

The online version of this article (10.1186/s13287-019-1205-1) contains supplementary material, which is available to authorized users.

Identification of potential key genes associated with osteosarcoma based on integrated bioinformatics analyses

Due to high rates of metastasis and poor clinical outcomes for patients, it is important to study the pathomechanisms of osteosarcoma. However, due to the fact that osteosarcoma shows significant interindividual variation and high heterogeneity, the identification of differentially expressed genes (DEGs) at the population level cannot answer many important questions related to osteosarcoma tumorigenesis. Therefore, a new strategy to identify dysregulated genes in osteosarcoma samples is required. The aim of this study was to improve our understanding of osteosarcoma pathogenesis by identifying genes with universal aberrant expression in osteosarcoma samples. Because the relative expression ordering of genes is stable in normal bone tissues but is disrupted in osteosarcoma tissues, we used the RankComp algorithm to identify DEGs in normal and osteosarcoma tissue samples. We then calculated the dysregulation frequency for each gene. Genes with deregulation frequencies above 80% were deemed to be universal DEGs. Next, coexpression, pathway enrichment, and protein-protein interaction network analyses were performed to characterize the functions of these genes. From 188 samples of osteosarcoma obtained from four datasets measured on different platforms, 51 universal DEGs were identified, including 4 universally upregulated genes and 47 universally downregulated genes. Genes that were differentially coexpressed with these universal DEGs were found to be enriched in 46 cancer-related pathways. In addition, functional and network analyses showed that genes with high dysregulation frequencies were involved in cancer-related functions. Thus, the commonly aberrant genes identified in osteosarcoma tissues may be important targets for osteosarcoma diagnosis and therapy.

A qualitative signature for early diagnosis of hepatocellular carcinoma based on relative expression orderings

BACKGROUND & AIMS

Currently, using biopsy specimens to confirm suspicious liver lesions of early hepatocellular carcinoma are not entirely reliable because of insufficient sampling amount and inaccurate sampling location. It is necessary to develop a signature to aid early hepatocellular carcinoma diagnosis using biopsy specimens even when the sampling location is inaccurate.

METHODS

Based on the within-sample relative expression orderings of gene pairs, we identified a simple qualitative signature to distinguish both hepatocellular carcinoma and adjacent non-tumour tissues from cirrhosis tissues of non-hepatocellular carcinoma patients.

RESULTS

A signature consisting of 19 gene pairs was identified in the training data sets and validated in 2 large collections of samples from biopsy and surgical resection specimens. For biopsy specimens, 95.7% of 141 hepatocellular carcinoma tissues and all (100%) of 108 cirrhosis tissues of non-hepatocellular carcinoma patients were correctly classified. Especially, all (100%) of 60 hepatocellular carcinoma adjacent normal tissues and 77.5% of 80 hepatocellular carcinoma adjacent cirrhosis tissues were classified to hepatocellular carcinoma. For surgical resection specimens, 99.7% of 733 hepatocellular carcinoma specimens were correctly classified to hepatocellular carcinoma, while 96.1% of 254 hepatocellular carcinoma adjacent cirrhosis tissues and 95.9% of 538 hepatocellular carcinoma adjacent normal tissues were classified to hepatocellular carcinoma. In contrast, 17.0% of 47 cirrhosis from non-hepatocellular carcinoma patients waiting for liver transplantation were classified to hepatocellular carcinoma, indicating that some patients with long-lasting cirrhosis could have already gained hepatocellular carcinoma characteristics.

CONCLUSIONS

The signature can distinguish both hepatocellular carcinoma tissues and tumour-adjacent tissues from cirrhosis tissues of non-hepatocellular carcinoma patients even using inaccurately sampled biopsy specimens, which can aid early diagnosis of hepatocellular carcinoma.

Transcriptional signatures for coupled predictions of stage II and III colorectal cancer metastasis and fluorouracil-based adjuvant chemotherapy benefit

The current study suggests that the identification of predictive signatures of fluorouracil (5-FU) response for stage II and III colorectal cancer (CRC) could be confounded by chemotherapy-irrelevant low relapse risk. Using the samples of patients with stage II and III CRC who were treated with curative surgery only, we identified a signature with which to predict chemotherapy-irrelevant relapse risk for patients after curative surgery. By applying this signature to the samples of patients with stage II and III CRC who were treated with 5-FU-based adjuvant chemotherapy (ACT) after surgery, we predicted the relapse risk if treated with surgery only. From high-risk samples, we further identified another signature with which to predict therapeutic benefit from 5-FU-based ACT. On the basis of the relative expression orderings of gene pairs, a postsurgery relapse risk signature that consisted of 44 gene pairs was developed and verified in 3 independent data sets. A 5-FU therapeutic benefit signature that consisted of 4 gene pairs was then developed to predict the response of 5-FU-based ACT for those patients with high relapse risk after curative surgery. The signature was verified in 4 independent datasets. For patients with stage II and III CRC, the coupled signatures can first identify patients with high relapse risk after curative surgery, then predict therapeutic benefit from 5-FU-based ACT.-Song, K., Guo, Y., Wang, X., Cai, H., Zheng, W., Li, N., Song, X., Ao, L., Guo, Z., Zhao, W. Transcriptional signatures for coupled predictions of stage II and III colorectal cancer metastasis and fluorouracil-based adjuvant chemotherapy benefit.

Individualized predictive signatures for 5-fluorouracil-based chemotherapy in right- and left-sided colon cancer

5‐Fluorouracil (5‐FU)‐based adjuvant chemotherapy (ACT) is widely used for the treatment of colon cancer. Colon cancers with different primary tumor locations are clinically and molecularly distinct, implied through their response to 5‐FU‐based ACT. In this work, using 69 and 133 samples of patients with stage II‐III right‐sided and left‐sided colon cancer (RCC and LCC) treated with post‐surgery 5‐FU‐based ACT, we preselected gene pairs whose relative expression orderings were significantly correlated with the disease‐free survival of patients by univariate Cox proportional hazards model. Then, from the identified prognostic‐related gene pairs, a forward‐stepwise selection algorithm was formulated to search for an optimal subset of gene pairs that resulted in the highest concordance index, referred to as the gene pair signature (GPS). We identified prognostic signatures, 3‐GPS and 5‐GPS, for predicting response to 5‐FU‐based ACT of patients with RCC and LCC, respectively, which were validated in independent datasets of GSE14333 and GSE72970. With the aid of the signatures, the transcriptional and genomic characteristics between the predicted responders and non‐responders were explored. Notably, both in RCC and LCC, the predicted responders to 5‐FU‐based ACT were characterized by hypermutation, whereas the predicted non‐responders were characterized by frequent copy number alternations. Finally, in comparison with the established relative expression ordering‐based signature, which was developed without considering the differences between RCC and LCC, the newly proposed signatures had a better predictive performance. In conclusion, 3‐GPS or 5‐GPS can robustly predict response to 5‐FU‐based ACT for patients with RCC or LCC, respectively, in an individual level.

Identifying differentially expressed genes from cross-site integrated data based on relative expression orderings

It is a basic task in high-throughput gene expression profiling studies to identify differentially expressed genes (DEGs) between two phenotypes. But the weakly differential expression signals between two phenotypes are hardly detectable with limited sample sizes. To solve this problem, many researchers tried to combine multiple independent datasets using meta-analysis or batch effect adjustment algorithms. However, these algorithms may distort true biological differences between two phenotypes and introduce unacceptable high false rates, as demonstrated in this study. These problems pose critical obstacles for analyzing the transcriptional data in The Cancer Genome Atlas where there are many small-scale batches of data. Previously, we developed RankComp to detect DEGs for individual disease samples through exploiting the incongruous relative expression orderings between two phenotypes and further improved it here to identify DEGs using multiple independent datasets. We demonstrated the improved RankComp can directly analyze integrated cross-site data to detect DEGs between two phenotypes without the need of batch effect adjustments. Its usage was illustrated in detecting weak differential expression signals of breast cancer drug-response data using combined datasets from multiple experiments.

Robust transcriptional tumor signatures applicable to both formalin-fixed paraffin-embedded and fresh-frozen samples

Formalin-fixed paraffin-embedded (FFPE) samples represent a valuable resource for clinical researches. However, FFPE samples are usually considered an unreliable source for gene expression analysis due to the partial RNA degradation. In this study, through comparing gene expression profiles between FFPE samples and paired fresh-frozen (FF) samples for three cancer types, we firstly showed that expression measurements of thousands of genes had at least two-fold change in FFPE samples compared with paired FF samples. Therefore, for a transcriptional signature based on risk scores summarized from the expression levels of the signature genes, the risk score thresholds trained from FFPE (or FF) samples could not be applied to FF (or FFPE) samples. On the other hand, we found that more than 90% of the relative expression orderings (REOs) of gene pairs in the FF samples were maintained in their paired FFPE samples and largely unaffected by the storage time. The result suggested that the REOs of gene pairs were highly robust against partial RNA degradation in FFPE samples. Finally, as a case study, we developed a REOs-based signature to distinguish liver cirrhosis from hepatocellular carcinoma (HCC) using FFPE samples. The signature was validated in four datasets of FFPE samples and eight datasets of FF samples. In conclusion, the valuable FFPE samples can be fully exploited to identify REOs-based diagnostic and prognostic signatures which could be robustly applicable to both FF samples and FFPE samples with degraded RNA.

Quantitative or qualitative transcriptional diagnostic signatures? A case study for colorectal cancer

Background

Due to experimental batch effects, the application of a quantitative transcriptional signature for disease diagnoses commonly requires inter-sample data normalization, which would be hardly applicable under common clinical settings. Many cancers might have qualitative differences with the non-cancer states in the gene expression pattern. Therefore, it is reasonable to explore the power of qualitative diagnostic signatures which are robust against experimental batch effects and other random factors.

Results

Firstly, using data of technical replicate samples from the MicroArray Quality Control (MAQC) project, we demonstrated that the low-throughput PCR-based technologies also exist large measurement variations for gene expression even when the samples were measured in the same test site. Then, we demonstrated the critical limitation of low stability for classifiers based on quantitative transcriptional signatures in applications to individual samples through a case study using a support vector machine and a naïve Bayesian classifier to discriminate colorectal cancer tissues from normal tissues. To address this problem, we identified a signature consisting of three gene pairs for discriminating colorectal cancer tissues from non-cancer (normal and inflammatory bowel disease) tissues based on within-sample relative expression orderings (REOs) of these gene pairs. The signature was well verified using 22 independent datasets measured by different microarray and RNA_seq platforms, obviating the need of inter-sample data normalization.

Conclusions

Subtle quantitative information of gene expression measurements tends to be unstable under current technical conditions, which will introduce uncertainty to clinical applications of the quantitative transcriptional diagnostic signatures. For diagnosis of disease states with qualitative transcriptional characteristics, the qualitative REO-based signatures could be robustly applied to individual samples measured by different platforms.

Electronic supplementary material

The online version of this article (10.1186/s12864-018-4446-y) contains supplementary material, which is available to authorized users.

Robust transcriptional signatures for low-input RNA samples based on relative expression orderings

Background

It is often difficult to obtain sufficient quantity of RNA molecules for gene expression profiling under many practical situations. Amplification from low-input samples may induce artificial signals.

Results

We compared the expression measurements of low-input mRNA samples, from 25 pg to 1000 pg mRNA, which were amplified and profiled by Smart-seq, DP-seq and CEL-seq techniques using the Illumina HiSeq 2000 platform, with those of the paired high-input (50 ng) mRNA samples. Even with 1000 pg mRNA input, we found that thousands of genes had at least 2 folds-change of expression levels in the low-input samples compared with the corresponding paired high-input samples. Consequently, a transcriptional signature based on quantitative expression values and determined from high-input RNA samples cannot be applied to low-input samples, and vice versa. In contrast, the within-sample relative expression orderings (REOs) of approximately 90% of all the gene pairs in the high-input samples were maintained in the paired low-input samples with 1000 pg input mRNA molecules. Similar results were observed in the low-input total RNA samples amplified and profiled by the Whole-Genome DASL technique using the Illumina HumanRef-8 v3.0 platform. As a proof of principle, we developed REOs-based signatures from high-input RNA samples for discriminating cancer tissues and showed that they can be robustly applied to low-input RNA samples.

Conclusions

REOs-based signatures determined from the high-input RNA samples can be robustly applied to samples profiled with the low-input RNA samples, as low as the 1000 pg and 250 pg input samples but no longer stable in samples with less than 250 pg RNA input to a certain degree.

Electronic supplementary material

The online version of this article (10.1186/s12864-017-4280-7) contains supplementary material, which is available to authorized users.

Evaluating hepatocellular carcinoma cell lines for tumour samples using within-sample relative expression orderings of genes

BACKGROUND & AIMS

Concerns are raised about the representativeness of cell lines for tumours due to the culture environment and misidentification. Liver is a major metastatic destination of many cancers, which might further confuse the origin of hepatocellular carcinoma cell lines. Therefore, it is of crucial importance to understand how well they can represent hepatocellular carcinoma.

METHODS

The HCC-specific gene pairs with highly stable relative expression orderings in more than 99% of hepatocellular carcinoma but with reversed relative expression orderings in at least 99% of one of the six types of cancer, colorectal carcinoma, breast carcinoma, non-small-cell lung cancer, gastric carcinoma, pancreatic carcinoma and ovarian carcinoma, were identified.

RESULTS

With the simple majority rule, the HCC-specific relative expression orderings from comparisons with colorectal carcinoma and breast carcinoma could exactly discriminate primary hepatocellular carcinoma samples from both primary colorectal carcinoma and breast carcinoma samples. Especially, they correctly classified more than 90% of liver metastatic samples from colorectal carcinoma and breast carcinoma to their original tumours. Finally, using these HCC-specific relative expression orderings from comparisons with six cancer types, we identified eight of 24 hepatocellular carcinoma cell lines in the Cancer Cell Line Encyclopedia (Huh-7, Huh-1, HepG2, Hep3B, JHH-5, JHH-7, C3A and Alexander cells) that are highly representative of hepatocellular carcinoma. Evaluated with a REOs-based prognostic signature for hepatocellular carcinoma, all these eight cell lines showed the same metastatic properties of the high-risk metastatic hepatocellular carcinoma tissues.

CONCLUSIONS

Caution should be taken for using hepatocellular carcinoma cell lines. Our results should be helpful to select proper hepatocellular carcinoma cell lines for biological experiments.

An individualized gene expression signature for prediction of lung adenocarcinoma metastases

Our laboratory previously reported an individual‐level signature consisting of nine gene pairs, named 9‐GPS. This signature was developed by training on microarray expression data and validated using three independent integrated microarray data sets, with samples of stage I non‐small‐cell lung cancer after complete surgical resection. In this study, we first validated the cross‐platform robustness of 9‐GPS by demonstrating that 9‐GPS could significantly stratify the overall survival of 213 stage I lung adenocarcinoma (LUAD) patients detected with RNA‐sequencing platform in The Cancer Genome Atlas (TCGA; log‐rank P = 0.0318, C‐index = 0.55). Applying 9‐GPS to all the 423 stage I‐IV LUAD samples in TCGA, the predicted high‐risk samples were significantly enriched with clinically diagnosed metastatic samples (Fisher's exact test, P = 0.0015). We further modified the voting rule of 9‐GPS and found that the modified 9‐GPS had a better performance in predicting metastasis states (Fisher's exact test, P < 0.0001). With the aid of the modified 9‐GPS for reclassifying the metastasis states of patients with LUAD, the reclassified metastatic samples presented clearer transcriptional and genomic characteristics compared to the reclassified nonmetastatic samples. Finally, regulator network analysis identified TP53 and IRF1 with frequent genomic aberrations in the reclassified metastatic samples, indicating their key roles in driving tumor metastasis. In conclusion, 9‐GPS is a robust signature for identifying early‐stage LUAD patients with potential occult metastasis. This occult metastasis prediction was associated with clear transcriptional and genomic characteristics as well as the clinical diagnoses.

Statistically controlled identification of differentially expressed genes in one-to-one cell line comparisons of the CMAP database for drug repositioning

BACKGROUND

The Connectivity Map (CMAP) database, an important public data source for drug repositioning, archives gene expression profiles from cancer cell lines treated with and without bioactive small molecules. However, there are only one or two technical replicates for each cell line under one treatment condition. For such small-scale data, current fold-changes-based methods lack statistical control in identifying differentially expressed genes (DEGs) in treated cells. Especially, one-to-one comparison may result in too many drug-irrelevant DEGs due to random experimental factors. To tackle this problem, CMAP adopts a pattern-matching strategy to build "connection" between disease signatures and gene expression changes associated with drug treatments. However, many drug-irrelevant genes may blur the "connection" if all the genes are used instead of pre-selected DEGs induced by drug treatments.

METHODS

We applied OneComp, a customized version of RankComp, to identify DEGs in such small-scale cell line datasets. For a cell line, a list of gene pairs with stable relative expression orderings (REOs) were identified in a large collection of control cell samples measured in different experiments and they formed the background stable REOs. When applying OneComp to a small-scale cell line dataset, the background stable REOs were customized by filtering out the gene pairs with reversal REOs in the control samples of the analyzed dataset.

RESULTS

In simulated data, the consistency scores of overlapping genes between DEGs identified by OneComp and SAM were all higher than 99%, while the consistency score of the DEGs solely identified by OneComp was 96.85% according to the observed expression difference method. The usefulness of OneComp was exemplified in drug repositioning by identifying phenformin and metformin related genes using small-scale cell line datasets which helped to support them as a potential anti-tumor drug for non-small-cell lung carcinoma, while the pattern-matching strategy adopted by CMAP missed the two connections. The implementation of OneComp is available at https://github.com/pathint/reoa .

CONCLUSIONS

OneComp performed well in both the simulated and real data. It is useful in drug repositioning studies by helping to find hidden "connections" between drugs and diseases.

Identifying CpG sites with different differential methylation frequencies in colorectal cancer tissues based on individualized differential methylation analysis

A big challenge to clinical diagnosis and therapy of colorectal cancer (CRC) is its extreme heterogeneity, and thus it would be of special importance if we could find common biomarkers besides subtype-specific biomarkers for CRC. Here, with DNA methylation data produced by different laboratories, we firstly revealed that the relative methylation-level orderings (RMOs) of CpG sites within colorectal normal tissues are highly stable but widely disrupted in the CRC tissues. This finding provides the basis for using the RankComp algorithm to identify differentially methylated (DM) CpG sites in every individual CRC sample through comparing the RMOs within the individual sample with the stable RMOs predetermined in normal tissues. For 75 CRC samples, RankComp detected averagely 4,062 DM CpG sites per sample and reached an average precision of 91.34% in terms that the hypermethylation or hypomethylation states of the DM CpG sites detected for each cancer sample were consistent with the observed differences between this cancer sample and its paired adjacent normal sample. Finally, we applied RankComp to identify DM CpG sites for each of the 268 CRC samples from The Cancer Genome Atlas and found 26 and 143 genes whose promoter regions included CpG sites that were hypermethylated and hypomethylated, respectively, in more than 95% of the 268 CRC samples. Individualized pathway analysis identified six pathways that were significantly enriched with DM genes in more than 90% of the CRC tissues. These universal DNA methylation biomarkers could be important diagnostic makers and therapy targets for CRC.

An individualised signature for predicting response with concordant survival benefit for lung adenocarcinoma patients receiving platinum-based chemotherapy

BACKGROUND

For lung adenocarcinoma (LUAD) patients receiving platinum-based adjuvant chemotherapy (ACT), predictive signatures extracted from survival data solely are not directly associated with platinum response. Another limitation of reported signatures, commonly based on risk scores summarised from gene expressions, is that they could not be applied directly to samples measured by different laboratories due to experimental batch effects.

METHODS

Using 60 samples of LUAD patients receiving platinum-based ACT in TCGA, we pre-selected gene pairs whose within-samples relative expression orderings (REOs) were significantly associated with both pathological response and 5-year survival, from which we selected an optimal signature whose within-samples REOs could identify responders with improved 5-year survival rate.

RESULTS

A predictive signature consisting of three gene pairs was developed. In an independent data set integrated from five small data sets, the predicted responders had a significantly higher 5-year survival rate than the predicted non-responders if and only if they received platinum-based ACT (log-rank P=0.0006). The predicted responders showed a 22% absolute benefit of platinum-based ACT in 5-year survival rate compared with untreated patients (log-rank P=0.0019).

CONCLUSIONS

The REO-based signature can individually predict response to platinum-based ACT with concordant survival benefit directly for LUAD samples measured by different laboratories.

Discriminating cancer-related and cancer-unrelated chemoradiation-response genes for locally advanced rectal cancers

For patients with locally advanced rectal cancer (LARC) treated with preoperation chemoradiation (pCRT), identifying differentially expressed (DE) genes between non-responders and responders is a common approach for investigating mechanisms of chemoradiation resistance. However, some of such DE genes might be irrelevant to cancer itself but simply reflect the pharmacokinetic differences of the normal tissues. In this study, we adopted the RankComp algorithm to identify DE genes for each of LARC sample compared with its own normal state. Then, we identified genes with significantly different deregulation frequencies between the non-responders and responders, defined as cancer-related pCRT-response genes. Pathway enrichment and protein-protein interaction analyses showed that these genes specifically and intensively interacted with currently known effective genes of pCRT, involving in DNA replication, cell cycle and DNA repair. In contrast, after excluding the cancer-related pCRT-response genes, the other DE genes between non-responders and responders were enriched in many pathways of drug and protein metabolisms and transports, and interacted with both the known effective genes and pharmacokinetic genes. Hence, these two types of DE genes should be distinguished for investigating mechanisms of pCRT response in LARCs.