Cooperative genomic alteration network reveals molecular classification across 12 major cancer types

Synergistic epistasis among cancer drivers can rescue early tumors from the accumulation of deleterious passengers

Epistasis among driver mutations is pervasive and explains relevant features of cancer, such as differential therapy response and convergence towards well-characterized molecular subtypes. Furthermore, a growing body of evidence suggests that tumor development could be hampered by the accumulation of slightly deleterious passenger mutations. In this work, we combined empirical epistasis networks, computer simulations, and mathematical models to explore how synergistic interactions among driver mutations affect cancer progression under the burden of slightly deleterious passengers. We found that epistasis plays a crucial role in tumor development by promoting the transformation of precancerous clones into rapidly growing tumors through a process that is analogous to evolutionary rescue. The triggering of epistasis-driven rescue is strongly dependent on the intensity of epistasis and could be a key rate-limiting step in many tumors, contributing to their unpredictability. As a result, central genes in cancer epistasis networks appear as key intervention targets for cancer therapy.

Small extrachromosomal circular DNAs as biomarkers for multi-cancer diagnosis and monitoring

BACKGROUND

Small extrachromosomal circular DNAs (eccDNAs) have the potential to be cancer biomarkers. However, the formation mechanisms and functions of small eccDNAs selected in carcinogenesis are not clear, and whether the small eccDNA profile in the plasma of cancer patients represents that in cancer tissues remains to be elucidated.

METHODS

A novel sequencing workflow based on the nanopore sequencing platform was used to sequence naturally existing full-length small eccDNAs in tissues and plasma collected from 25 cancer patients (including prostate cancer, hepatocellular carcinoma and colorectal cancer), and from an independent validation cohort (including 7 cancer plasma and 14 healthy plasma).

RESULTS

Compared with those in non-cancer tissues, small eccDNAs detected in cancer tissues had a significantly larger number and size (P = 0.040 and 2.2e-16, respectively), along with more even distribution and different formation mechanisms. Although small eccDNAs had different general characteristics and genomic annotation between cancer tissues and the paired plasma, they had similar formation mechanisms and cancer-related functions. Small eccDNAs originated from some specific genes had great multi-cancer diagnostic value in tissues (AUC ≥ 0.8) and plasma (AUC > 0.9), especially increasing the accuracy of multi-cancer prediction of CEA/CA19-9 levels. The high multi-cancer diagnostic value of small eccDNAs originated from ALK&ETV6 could be extrapolated from tissues (AUC = 0.804) to plasma and showed high positive predictive value (100%) and negative predictive value (82.35%) in a validation cohort.

CONCLUSIONS

As independent and stable circular DNA molecules, small eccDNAs in both tissues and plasma can be used as ideal biomarkers for cost-effective multi-cancer diagnosis and monitoring.

Pervasive conditional selection of driver mutations and modular epistasis networks in cancer

Cancer driver mutations often display mutual exclusion or co-occurrence, underscoring the key role of epistasis in carcinogenesis. However, estimating the magnitude of epistasis and quantifying its effect on tumor evolution remains a challenge. We develop a method (Coselens) to quantify conditional selection on the excess of nonsynonymous substitutions in cancer genes. Coselens infers the number of drivers per gene in different partitions of a cancer genomics dataset using covariance-based mutation models and determines whether coding mutations in a gene affect selection for drivers in any other gene. Using Coselens, we identify 296 conditionally selected gene pairs across 16 cancer types in the TCGA dataset. Conditional selection affects 25%-50% of driver substitutions in tumors with >2 drivers. Conditionally co-selected genes form modular networks, whose structures challenge the traditional interpretation of within-pathway mutual exclusivity and across-pathway synergy, suggesting a more complex scenario where gene-specific across-pathway epistasis shapes differentiated cancer subtypes.

Pediatric Pan-Central Nervous System Tumor Methylome Analyses Reveal Immune-Related LncRNAs

Pediatric central nervous system (CNS) tumors are the second most common cancer diagnosis among children. Long noncoding RNAs (lncRNAs) emerge as critical regulators of gene expression, and they play fundamental roles in immune regulation. However, knowledge on epigenetic changes in lncRNAs in diverse types of pediatric CNS tumors is lacking. Here, we integrated the DNA methylation profiles of 2,257 pediatric CNS tumors across 61 subtypes with lncRNA annotations and presented the epigenetically regulated landscape of lncRNAs. We revealed the prevalent lncRNA methylation heterogeneity across pediatric pan-CNS tumors. Based on lncRNA methylation profiles, we refined 14 lncRNA methylation clusters with distinct immune microenvironment patterns. Moreover, we found that lncRNA methylations were significantly correlated with immune cell infiltrations in diverse tumor subtypes. Immune-related lncRNAs were further identified by investigating their correlation with immune cell infiltrations and potentially regulated target genes. LncRNA with methylation perturbations potentially regulate the genes in immune-related pathways. We finally identified several candidate immune-related lncRNA biomarkers (i.e., SSTR5-AS1, CNTN4-AS1, and OSTM1-AS1) in pediatric cancer for further functional validation. In summary, our study represents a comprehensive repertoire of epigenetically regulated immune-related lncRNAs in pediatric pan-CNS tumors, and will facilitate the development of immunotherapeutic targets.

English Speech Recognition System Model Based on Computer-Aided Function and Neural Network Algorithm

With the economic globalization continuous growth of China's socioeconomic level tends to be internationalized, China's attention to English has been significantly improved. However, the domestic English teaching level is limited, so it is impossible to correct students' English pronunciation and make a reasonable evaluation at all times so that oral training has certain disadvantages. However, the computer-aided language learning system at home and abroad focuses on the practice of words and grammar, and the evaluation indicators are less and not comprehensive. In view of the complexity of English pronunciation changes, traditional speech recognition is difficult to recognize speech speed and improve its accuracy. Furthermore, to strengthen the English pronunciation of domestic students, a nonlinear network structure is studied in depth to simulate the human brain to analyze a model of speech recognition is established Mel frequency cepstrum characteristic parameters of human ear model and deep belief network. In this paper, the traditional computer pronunciation evaluation method is improved in an all-round way, and a set of high-quality speech recognition system of speech recognition method is constructed. Aiming at the above problems, it takes the students as the research, which proves that the method adopted in this paper can give the learners accurate pronunciation quality analysis report and guidance and correct their intonation and improve the learning effect, and the experimental data verify that the improved speech recognition system model recognition ability is higher than the traditional model.

Revealing the contribution of somatic gene mutations to shaping tumor immune microenvironment

Interaction between tumor cells and immune cells determined highly heterogeneous microenvironments across patients, leading to substantial variation in clinical benefits from immunotherapy. Somatic gene mutations were found not only to elicit adaptive immunity but also to influence the composition of tumor immune microenvironment and various processes of antitumor immunity. However, due to an incomplete view of associations between gene mutations and immunophenotypes, how tumor cells shape the immune microenvironment and further determine the clinical benefit of immunotherapy is still unclear. To address this, we proposed a computational approach, inference of mutation effect on immunophenotype by integrated gene set enrichment analysis (MEIGSEA), for tracing back the genomic factor responsible for differences in immunophenotypes. MEIGSEA was demonstrated to accurately identify the previous confirmed immune-associated gene mutations, and systematic evaluation in simulation data further supported its performance. We used MEIGSEA to investigate the influence of driver gene mutations on the infiltration of 22 immune cell types across 19 cancers from The Cancer Genome Atlas. The top associated gene mutations with infiltration of CD8 T cells, such as CASP8, KRAS and EGFR, also showed extensive impact on other immune components; meanwhile, immune effector cells shared critical gene mutations that collaboratively contribute to shaping distinct tumor immune microenvironment. Furthermore, we highlighted the predictive capacity of gene mutations that are positively associated with CD8 T cells for the clinical benefit of immunotherapy. Taken together, we present a computational framework to help illustrate the potential of somatic gene mutations in shaping the tumor immune microenvironment.

Identifying Key Somatic Copy Number Alterations Driving Dysregulation of Cancer Hallmarks in Lower-Grade Glioma

Somatic copy-number alterations (SCNAs) are major contributors to cancer development that are pervasive and highly heterogeneous in human cancers. However, the driver roles of SCNAs in cancer are insufficiently characterized. We combined network propagation and linear regression models to design an integrative strategy to identify driver SCNAs and dissect the functional roles of SCNAs by integrating profiles of copy number and gene expression in lower-grade glioma (LGG). We applied our strategy to 511 LGG patients and identified 98 driver genes that dysregulated 29 cancer hallmark signatures, forming 143 active gene-hallmark pairs. We found that these active gene-hallmark pairs could stratify LGG patients into four subtypes with significantly different survival times. The two new subtypes with similar poorest prognoses were driven by two different gene sets (one including EGFR, CDKN2A, CDKN2B, INFA8, and INFA5, and the other including CDK4, AVIL, and DTX3), respectively. The SCNAs of the two gene sets could disorder the same cancer hallmark signature in a mutually exclusive manner (including E2F_TARGETS and G2M_CHECKPOINT). Compared with previous methods, our strategy could not only capture the known cancer genes and directly dissect the functional roles of their SCNAs in LGG, but also discover the functions of new driver genes in LGG, such as IFNA5, IFNA8, and DTX3. Additionally, our method can be applied to a variety of cancer types to explore the pathogenesis of driver SCNAs and improve the treatment and diagnosis of cancer.

Transcriptomic heterogeneity of driver gene mutations reveals novel mutual exclusivity and improves exploration of functional associations

Background

Lung adenocarcinoma (LUAD), as the most common subtype of lung cancer, is the leading cause of cancer deaths in the world. The accumulation of driver gene mutations enables cancer cells to gradually acquire growth advantage. Therefore, it is important to understand the functions and interactions of driver gene mutations in cancer progression.

Methods

We obtained gene mutation data and gene expression profile of 506 LUAD tumors from The Cancer Genome Atlas (TCGA). The subtypes of tumors with driver gene mutations were identified by consensus cluster analysis.

Results

We found 21 significantly mutually exclusive pairs consisting of 20 genes among 506 LUAD patients. Because of the increased transcriptomic heterogeneity of mutations, we identified subtypes among tumors with non‐silent mutations in driver genes. There were 494 mutually exclusive pairs found among driver gene mutations within different subtypes. Furthermore, we identified functions of mutually exclusive pairs based on the hypothesis of functional redundancy of mutual exclusivity. These mutually exclusive pairs were significantly enriched in nuclear division and humoral immune response, which played crucial roles in cancer initiation and progression. We also found 79 mutually exclusive triples among subtypes of tumors with driver gene mutations, which were key roles in cell motility and cellular chemical homeostasis. In addition, two mutually exclusive triples and one mutually exclusive triple were associated with the overall survival and disease‐specific survival of LUAD patients, respectively.

Conclusions

We revealed novel mutual exclusivity and generated a comprehensive functional landscape of driver gene mutations, which could offer a new perspective to understand the mechanisms of cancer development and identify potential biomarkers for LUAD therapy.

Epigenetic dysregulation of immune-related pathways in cancer: bioinformatics tools and visualization

Cancer immune evasion is one of the hallmarks of carcinogenesis. Cancer cells employ multiple mechanisms to avoid immune recognition and suppress antitumor immune responses. Recently, accumulating evidence has indicated that immune-related pathways are epigenetically dysregulated in cancer. Most importantly, the epigenetic footprint of immune-related pathways is associated with the patient outcome, underscoring the crucial need to understand this process. In this review, we summarize the current evidence for epigenetic regulation of immune-related pathways in cancer and describe bioinformatics tools, informative visualization techniques, and resources to help decipher the cancer epigenome.

Abnormal patterns of genomic chemical modification help tumors elude immunological destruction, but sophisticated computational tools could help identify and overcome these survival mechanisms. Immunotherapy can be a potent weapon against cancer, but many tumors evolve the ability to protect themselves by subduing the immune response. Sungjune Kim and colleagues at the Moffitt Cancer Center, Tampa, USA, have reviewed efforts to study how chemical alterations to DNA that affect gene expression contribute to this process. Considerable evidence indicates a role for a modification called methylation in this immune evasion, and researchers now have access to vast repositories of tumor-specific gene methylation profiles. The authors describe these data resources, and highlight some of the software tools that are helping oncologists to identify patterns in the data that might lead to better therapies.

Identification of novel prognostic biomarkers by integrating multi-omics data in gastric cancer

BACKGROUND

Gastric cancer is a fatal gastrointestinal cancer with high morbidity and poor prognosis. The dismal 5-year survival rate warrants reliable biomarkers to assess and improve the prognosis of gastric cancer. Distinguishing driver mutations that are required for the cancer phenotype from passenger mutations poses a formidable challenge for cancer genomics.

METHODS

We integrated the multi-omics data of 293 primary gastric cancer patients from The Cancer Genome Atlas (TCGA) to identify key driver genes by establishing a prognostic model of the patients. Analyzing both copy number alteration and somatic mutation data helped us to comprehensively reveal molecular markers of genomic variation. Integrating the transcription level of genes provided a unique perspective for us to discover dysregulated factors in transcriptional regulation.

RESULTS

We comprehensively identified 31 molecular markers of genomic variation. For instance, the copy number alteration of WASHC5 (also known as KIAA0196) frequently occurred in gastric cancer patients, which cannot be discovered using traditional methods based on significant mutations. Furthermore, we revealed that several dysregulation factors played a hub regulatory role in the process of biological metabolism based on dysregulation networks. Cancer hallmark and functional enrichment analysis showed that these key driver (KD) genes played a vital role in regulating programmed cell death. The drug response patterns and transcriptional signatures of KD genes reflected their clinical application value.

CONCLUSIONS

These findings indicated that KD genes could serve as novel prognostic biomarkers for further research on the pathogenesis of gastric cancers. Our study elucidated a multidimensional and comprehensive genomic landscape and highlighted the molecular complexity of GC.

AtMAD: Arabidopsis thaliana multi-omics association database

Integration analysis of multi-omics data provides a comprehensive landscape for understanding biological systems and mechanisms. The abundance of high-quality multi-omics data (genomics, transcriptomics, methylomics and phenomics) for the model organism Arabidopsis thaliana enables scientists to study the genetic mechanism of many biological processes. However, no resource is available to provide comprehensive and systematic multi-omics associations for Arabidopsis. Here, we developed an Arabidopsis thaliana Multi-omics Association Database (AtMAD, http://www.megabionet.org/atmad), a public repository for large-scale measurements of associations between genome, transcriptome, methylome, pathway and phenotype in Arabidopsis, designed for facilitating identification of eQTL, emQTL, Pathway-mQTL, Phenotype-pathway, GWAS, TWAS and EWAS. Candidate variants/methylations/genes were identified in AtMAD for specific phenotypes or biological processes, many of them are supported by experimental evidence. Based on the multi-omics association strategy, we have identified 11 796 cis-eQTLs and 10 119 trans-eQTLs. Among them, 68 837 environment-eQTL associations and 149 622 GWAS-eQTL associations were identified and stored in AtMAD. For expression–methylation quantitative trait loci (emQTL), we identified 265 776 emQTLs and 122 344 pathway-mQTLs. For TWAS and EWAS, we obtained 62 754 significant phenotype-gene associations and 3 993 379 significant phenotype-methylation associations, respectively. Overall, the multi-omics associated network in AtMAD will provide new insights into exploring biological mechanisms of plants at multi-omics levels.

Genome-wide methylation patterns predict clinical benefit of immunotherapy in lung cancer

Background

It is crucial to unravel molecular determinants of responses to immune checkpoint blockade (ICB) therapy because only a small subset of advanced non-small cell lung cancer (NSCLC) patients responds to ICB therapy. Previous studies were concentrated on genomic and transcriptomic markers (e.g., mutation burden and immune gene expression). However, these markers are not sufficient to accurately predict a response to ICB therapy.

Results

Here, we analyzed DNA methylomes of 141 advanced NSCLC samples subjected to ICB therapy (i.e., anti-programmed death-1) from two independent cohorts (60 and 81 patients from our and IDIBELL cohorts). Integrative analysis of patients with matched transcriptome data in our cohort (n = 28) at pathway level revealed significant overlaps between promoter hypermethylation and transcriptional repression in nonresponders relative to responders. Fifteen immune-related pathways, including interferon signaling, were identified to be enriched for both hypermethylation and repression. We built a reliable prognostic risk model based on eight genes using LASSO model and successfully validated the model in independent cohorts. Furthermore, we found 30 survival-associated molecular interaction networks, in which two or three hypermethylated genes showed significant mutual exclusion across nonresponders.

Conclusions

Our study demonstrates that methylation patterns can provide insight into molecular determinants underlying the clinical benefit of ICB therapy.

Identification of a Six-lncRNA Signature With Prognostic Value for Breast Cancer Patients

Breast cancer (BRCA) is the most common cancer and a major cause of death in women. Long non-coding RNAs (lncRNAs) are emerging as key regulators and have been implicated in carcinogenesis and prognosis. In this study, we aimed to develop a lncRNA signature of BRCA patients to improve risk stratification. In the training cohort (GSE21653, n = 232), 17 lncRNAs were identified by univariate Cox proportional hazards regression, which were significantly associated with patients’ survival. The least absolute shrinkage and selection operator-penalized Cox proportional hazards regression analysis was used to identify a six-lncRNA signature. According to the median of the signature risk score, patients were divided into a high-risk group and a low-risk group with significant disease-free survival differences in the training cohort. A similar phenomenon was observed in validation cohorts (GSE42568, n = 101; GSE20711, n = 87). The six-lncRNA signature remained as independent prognostic factors after adjusting for clinical factors in these two cohorts. Furthermore, this signature significantly predicted the survival of grade III patients and estrogen receptor-positive patients. Furthermore, in another cohort (GSE19615, n = 115), the low-risk patients that were treated with tamoxifen therapy had longer disease-free survival than those who underwent no therapy. Overall, the six-lncRNA signature can be a potential prognostic tool used to predict disease-free survival of patients and to predict the benefits of tamoxifen treatment in BRCA, which will be helpful in guiding individualized treatments for BRCA patients.

DNMIVD: DNA methylation interactive visualization database

Aberrant DNA methylation plays an important role in cancer progression. However, no resource has been available that comprehensively provides DNA methylation-based diagnostic and prognostic models, expression–methylation quantitative trait loci (emQTL), pathway activity-methylation quantitative trait loci (pathway-meQTL), differentially variable and differentially methylated CpGs, and survival analysis, as well as functional epigenetic modules for different cancers. These provide valuable information for researchers to explore DNA methylation profiles from different aspects in cancer. To this end, we constructed a user-friendly database named DNA Methylation Interactive Visualization Database (DNMIVD), which comprehensively provides the following important resources: (i) diagnostic and prognostic models based on DNA methylation for multiple cancer types of The Cancer Genome Atlas (TCGA); (ii) meQTL, emQTL and pathway-meQTL for diverse cancers; (iii) Functional Epigenetic Modules (FEM) constructed from Protein-Protein Interactions (PPI) and Co-Occurrence and Mutual Exclusive (COME) network by integrating DNA methylation and gene expression data of TCGA cancers; (iv) differentially variable and differentially methylated CpGs and differentially methylated genes as well as related enhancer information; (v) correlations between methylation of gene promoter and corresponding gene expression and (vi) patient survival-associated CpGs and genes with different endpoints. DNMIVD is freely available at http://www.unimd.org/dnmivd/. We believe that DNMIVD can facilitate research of diverse cancers.

Co-occurrence and Mutual Exclusivity Analysis of DNA Methylation Reveals Distinct Subtypes in Multiple Cancers

Co-occurrence and mutual exclusivity (COME) of DNA methylation refer to two or more genes that tend to be positively or negatively correlated in DNA methylation among different samples. Although COME of gene mutations in pan-cancer have been well explored, little is known about the COME of DNA methylation in pan-cancer. Here, we systematically explored the COME of DNA methylation profile in diverse human cancer. A total of 5,128,332 COME events were identified in 14 main cancers types in The Cancer Genome Atlas (TCGA). We also identified functional epigenetic modules of the zinc finger gene family in six cancer types by integrating the gene expression and DNA methylation data and the frequently occurred COME network. Interestingly, most of the genes in those functional epigenetic modules are epigenetically repressed. Strikingly, those frequently occurred COME events could be used to classify the patients into several subtypes with significant different clinical outcomes in six cancers as well as pan-cancer (p-value ≤ = 0.05). Moreover, we observed significant associations between different COME subtypes and clinical features (e.g., age, gender, histological type, neoplasm histologic grade, and pathologic stage) in distinct cancers. Taken together, we identified millions of COME events of DNA methylation in pan-cancer and detected functional epigenetic COME events that could separate tumor patients into different subtypes, which may benefit the diagnosis and prognosis of pan-cancer.

The correlation and role analysis of COL4A1 and COL4A2 in hepatocarcinogenesis

Liver fibrosis biomarker, Type IV collagen, may function as hepatocarcinogenesis niche. However, among the six isoforms, the isoforms providing tumor microenvironment and their regulatory network are still unclarified. Based on bioinformatics analysis of hundreds of HCC transcriptome datasets from public databases, we found that COL4A1/2 expressions were significantly correlated with hepatocarcinogenesis, progression, and prognosis. The expressions of COL4A1/2 were significantly upregulated in the preneoplastic and HCC tissues compared with normal tissues. Moreover, the overexpression of COL4A2 was highly correlated with shorter progression-free survival in HCC patients. Bioinformatics analysis also generates an interactive regulatory network in which COL4A1/2 directly binding to integrin alpha-2/beta-1 initiates a sequentially and complicated signaling transduction, to accelerate cell cycle and promote tumorigenesis. Among those pathways, the PI3K-Akt pathway is significantly enriched in cooperative mutations and correlation analysis. This suggests that the key activated signaling is PI3K-Akt pathway which severing as the centerline linked with other pathways (Wnt and MAPK signaling) and cell behaviors signaling (cell cycle control and cytoskeleton change). Switching extracellular matrix collagen isoform may establish pro-tumorigenic and metastatic niches. The findings of COL4A1/2 and related signaling networks are valuable to be further investigated that may provide druggable targets for HCC intervention.

Identifying subpathway signatures for individualized anticancer drug response by integrating multi-omics data

Background

Individualized drug response prediction is vital for achieving personalized treatment of cancer and moving precision medicine forward. Large-scale multi-omics profiles provide unprecedented opportunities for precision cancer therapy.

Methods

In this study, we propose a pipeline to identify subpathway signatures for anticancer drug response of individuals by integrating the comprehensive contributions of multiple genetic and epigenetic (gene expression, copy number variation and DNA methylation) alterations.

Results

Totally, 46 subpathway signatures associated with individual responses to different anticancer drugs were identified based on five cancer-drug response datasets. We have validated the reliability of subpathway signatures in two independent datasets. Furthermore, we also demonstrated these multi-omics subpathway signatures could significantly improve the performance of anticancer drug response prediction. In-depth analysis of these 46 subpathway signatures uncovered the essential roles of three omics types and the functional associations underlying different anticancer drug responses. Patient stratification based on subpathway signatures involved in anticancer drug response identified subtypes with different clinical outcomes, implying their potential roles as prognostic biomarkers. In addition, a landscape of subpathways associated with cellular responses to 191 anticancer drugs from CellMiner was provided and the mechanism similarity of drug action was accurately unclosed based on these subpathways. Finally, we constructed a user-friendly web interface-CancerDAP (http://bio-bigdata.hrbmu.edu.cn/CancerDAP/) available to explore 2751 subpathways relevant with 191 anticancer drugs response.

Conclusions

Taken together, our study identified and systematically characterized subpathway signatures for individualized anticancer drug response prediction, which may promote the precise treatment of cancer and the study for molecular mechanisms of drug actions.

Electronic supplementary material

The online version of this article (10.1186/s12967-019-2010-4) contains supplementary material, which is available to authorized users.

SEECancer: a resource for somatic events in evolution of cancer genome

Cancer cells progressively evolve from a premalignant to a malignant state, which is driven by accumulating somatic alterations that confer normal cells a fitness advantage. Improvements in high-throughput sequencing techniques have led to an increase in construction of tumor phylogenetics and identification of somatic driver events that specifically occurred in different tumor progression stages. Here, we developed the SEECancer database (http://biocc.hrbmu.edu.cn/SEECancer), which aims to present the comprehensive cancer evolutionary stage-specific somatic events (including early-specific, late-specific, relapse-specific, metastasis-specific, drug-resistant and drug-induced genomic events) and their temporal orders. By manually curating over 10 000 published articles, 1231 evolutionary stage-specific genomic events and 5772 temporal orders involving 82 human cancers and 23 tissue origins were collected and deposited in the SEECancer database. Each entry contains the somatic event, evolutionary stage, cancer type, detection approach and relevant evidence. SEECancer provides a user-friendly interface for browsing, searching and downloading evolutionary stage-specific somatic events and temporal relationships in various cancers. With increasing attention on cancer genome evolution, the necessary information in SEECancer will facilitate understanding of cancer etiology and development of evolutionary therapeutics, and help clinicians to discover biomarkers for monitoring tumor progression.

Identifying mutual exclusivity across cancer genomes: computational approaches to discover genetic interaction and reveal tumor vulnerability

Systematic sequencing of cancer genomes has revealed prevalent heterogeneity, with patients harboring various combinatorial patterns of genetic alteration. In particular, a phenomenon that a group of genes exhibits mutually exclusive patterns has been widespread across cancers, covering a broad spectrum of crucial cancer pathways. Recently, there is considerable evidence showing that, mutual exclusivity reflects alternative functions in tumor initiation and progression, or suggests adverse effects of their concurrence. Given its importance, numerous computational approaches have been proposed to study mutual exclusivity using genomic profiles alone, or by integrating networks and phenotypes. Some of them have been routinely used to explore genetic associations, which lead to a deeper understanding of carcinogenic mechanisms and reveals unexpected tumor vulnerabilities. Here, we present an overview of mutual exclusivity from the perspective of cancer genome. We describe the common hypothesis underlying mutual exclusivity, summarize the strategies for the identification of significant mutually exclusive patterns, compare the performance of representative algorithms from simulated data sets and discuss their common confounders.

Integrative analysis identifies potential DNA methylation biomarkers for pan-cancer diagnosis and prognosis

DNA methylation status is closely associated with diverse diseases, and is generally more stable than gene expression, thus abnormal DNA methylation could be important biomarkers for tumor diagnosis, treatment and prognosis. However, the signatures regarding DNA methylation changes for pan-cancer diagnosis and prognosis are less explored. Here we systematically analyzed the genome-wide DNA methylation patterns in diverse TCGA cancers with machine learning. We identified seven CpG sites that could effectively discriminate tumor samples from adjacent normal tissue samples for 12 main cancers of TCGA (1216 samples, AUC > 0.99). Those seven potential diagnostic biomarkers were further validated in the other 9 different TCGA cancers and 4 independent datasets (AUC > 0.92). Three out of the seven CpG sites were correlated with cell division, DNA replication and cell cycle. We also identified 12 CpG sites that can effectively distinguish 26 different cancers (7605 samples), and the result was repeatable in independent datasets as well as two disparate tumors with metastases (micro-average AUC > 0.89). Furthermore, a series of potential signatures that could significantly predict the prognosis of tumor patients for 7 different cancer were identified via survival analysis (p-value < 1e-4). Collectively, DNA methylation patterns vary greatly between tumor and adjacent normal tissues, as well as among different types of cancers. Our identified signatures may aid the decision of clinical diagnosis and prognosis for pan-cancer and the potential cancer-specific biomarkers could be used to predict the primary site of metastatic breast and prostate cancers.

A pan-cancer atlas of cancer hallmark-associated candidate driver lncRNAs

Substantial cancer genome sequencing efforts have discovered many important driver genes contributing to tumorigenesis. However, very little is known about the genetic alterations of long non‐coding RNAs (lncRNAs) in cancer. Thus, there is a need for systematic surveys of driver lncRNAs. Through integrative analysis of 5918 tumors across 11 cancer types, we revealed that lncRNAs have undergone dramatic genomic alterations, many of which are mutually exclusive with well‐known cancer genes. Using the hypothesis of functional redundancy of mutual exclusivity, we developed a computational framework to identify driver lncRNAs associated with different cancer hallmarks. Applying it to pan‐cancer data, we identified 378 candidate driver lncRNAs whose genomic features highly resemble the known cancer driver genes (e.g. high conservation and early replication). We further validated the candidate driver lncRNAs involved in ‘Tissue Invasion and Metastasis’ in lung adenocarcinoma and breast cancer, and also highlighted their potential roles in improving clinical outcomes. In summary, we have generated a comprehensive landscape of cancer candidate driver lncRNAs that could act as a starting point for future functional explorations, as well as the identification of biomarkers and lncRNA‐based target therapy.

Identifying prognostic signature in ovarian cancer using DirGenerank

Identifying the prognostic genes in cancer is essential not only for the treatment of cancer patients, but also for drug discovery. However, it's still a big challenge to select the prognostic genes that can distinguish the risk of cancer patients across various data sets because of tumor heterogeneity. In this situation, the selected genes whose expression levels are statistically related to prognostic risks may be passengers. In this paper, based on gene expression data and prognostic data of ovarian cancer patients, we used conditional mutual information to construct gene dependency network in which the nodes (genes) with more out-degrees have more chances to be the modulators of cancer prognosis. After that, we proposed DirGenerank (Generank in direct netowrk) algorithm, which concerns both the gene dependency network and genes' correlations to prognostic risks, to identify the gene signature that can predict the prognostic risks of ovarian cancer patients. Using ovarian cancer data set from TCGA (The Cancer Genome Atlas) as training data set, 40 genes with the highest importance were selected as prognostic signature. Survival analysis of these patients divided by the prognostic signature in testing data set and four independent data sets showed the signature can distinguish the prognostic risks of cancer patients significantly. Enrichment analysis of the signature with curated cancer genes and the drugs selected by CMAP showed the genes in the signature may be drug targets for therapy. In summary, we have proposed a useful pipeline to identify prognostic genes of cancer patients.