Cancer Mutation Data Research Articles

Construction of a cuproptosis‑related lncRNA signature to predict biochemical recurrence of prostate cancer.

Biochemical recurrence (BCR) is common in prostate cancer (PCa), and patients with BCR usually have a poor prognosis. Cuproptosis is a unique type of cell death, and copper homeostasis is crucial to the occurrence and development of malignancies. The present study aimed to explore the prognostic value of cuproptosis-related long non-coding RNAs (lncRNAs; CRLs) in PCa and to develop a predictive signature for forecasting BCR in patients with PCa. Using The Cancer Genome Atlas database, transcriptomic, mutation and clinical data were collected from patients with PCa. A total of 121 CRLs were identified using Pearson's correlation coefficient. Subsequently, a 6-CRL signature consisting of AC087276.2, CNNM3-DT, AC090198.1, AC138207.5, METTL14-DT and LINC01515 was created to predict the BCR of patients with PCa through Cox and least absolute shrinkage and selection operator regression analyses. Kaplan-Meier curve analysis demonstrated that high-risk patients had a low BCR-free survival rate. In addition, there was a substantial difference between the high- and low-risk groups in the immune microenvironment, immune therapy, drug sensitivity and tumor mutational burden. A nomogram integrating the Gleason score, 6-CRL signature and clinical T-stage was established and evaluated. Finally, the expression of signature lncRNAs in PCa cells was verified through reverse transcription-quantitative PCR. In conclusion, the 6-CRL signature may be a potential tool for making predictions regarding BCR in patients with PCa, and the prognostic nomogram may be considered a practical tool for clinical decision-making.

Differential expression of DMD transcripts as a novel prognostic biomarker in histologically diverse mesotheliomas.

The identification of prognostic biomarkers is crucial for guiding treatment strategies in mesothelioma patients. The Duchenne muscular dystrophy (DMD) gene and its specific transcripts have been associated with patient survival in various tumours. In this study, we aimed to investigate the prognostic potential of DMD gene expression and its transcripts in mesothelioma patients. We analysed The Cancer Genome Atlas (TCGA) mesothelioma RNAseq, mutation, and clinical data to assess the association between DMD gene expression and its transcripts (Dp427, Dp71 splice variants) and mesothelioma survival. We also evaluated the specific Dp71 transcript as a unique prognostic biomarker across mesothelioma subtypes. Additionally, we performed differential gene expression analysis between high and low DMD gene/transcript expression groups. The analysis included 57 epithelioid, 23 biphasic, two sarcomatoid, and five not otherwise specified (NOS) histological subtypes of mesothelioma samples. Univariate analysis revealed that high expression of the DMD gene and its Dp71 transcript was significantly associated with shorter survival in mesothelioma patients (P=0.003 and P<0.001, respectively). In a multivariate analysis, the association between Dp71 expression and survival remained significant [hazard ratio (HR) 2.29, 95% confidence interval (CI): 1.24-4.23, P=0.008] across all mesothelioma patients, and also among patients with mesotheliomas without deep CDKN2A deletions (HR 3.58, 95% CI: 1.31-9.80, P=0.01). Pathway analysis revealed enrichment of cell cycle (P=3.01×10-4) and homologous recombination (P=0.01) pathways in differentially expressed genes (DEGs) between high and low Dp71 groups. Furthermore, there were correlations between Dp71 transcript expression and tumour microenvironment (TME) cells, including a weak positive correlation with macrophages (R=0.32, P=0.002) specifically M2 macrophages (R=0.34, P=0.001). Our findings indicate that the differential expression of specific DMD transcripts is associated with poor survival in mesothelioma patients. The specific Dp71 transcript can serve as a potential biomarker for predicting patient survival in diverse histological subtypes of mesothelioma. Further studies are needed to understand the role of specific dystrophin transcripts in cancer and TME cells, and their implications in the pathogenesis and progression of mesothelioma. Identifying patients at risk of poor survival based on DMD transcript expression can guide treatment strategies in mesothelioma, informing decisions regarding treatment intensity, follow-up schedules, eligibility for clinical trials, and ultimately, end-of-life care planning.

Prediction of metabolites associated with somatic mutations in cancers by using genome-scale metabolic models and mutation data

BackgroundOncometabolites, often generated as a result of a gene mutation, show pro-oncogenic function when abnormally accumulated in cancer cells. Identification of such mutation-associated metabolites will facilitate developing treatment strategies for cancers, but is challenging due to the large number of metabolites in a cell and the presence of multiple genes associated with cancer development.ResultsHere we report the development of a computational workflow that predicts metabolite-gene-pathway sets. Metabolite-gene-pathway sets present metabolites and metabolic pathways significantly associated with specific somatic mutations in cancers. The computational workflow uses both cancer patient-specific genome-scale metabolic models (GEMs) and mutation data to generate metabolite-gene-pathway sets. A GEM is a computational model that predicts reaction fluxes at a genome scale and can be constructed in a cell-specific manner by using omics data. The computational workflow is first validated by comparing the resulting metabolite-gene pairs with multi-omics data (i.e., mutation data, RNA-seq data, and metabolome data) from acute myeloid leukemia and renal cell carcinoma samples collected in this study. The computational workflow is further validated by evaluating the metabolite-gene-pathway sets predicted for 18 cancer types, by using RNA-seq data publicly available, in comparison with the reported studies. Therapeutic potential of the resulting metabolite-gene-pathway sets is also discussed.ConclusionsValidation of the metabolite-gene-pathway set-predicting computational workflow indicates that a decent number of metabolites and metabolic pathways appear to be significantly associated with specific somatic mutations. The computational workflow and the resulting metabolite-gene-pathway sets will help identify novel oncometabolites and also suggest cancer treatment strategies.

Standing in the canine precision medicine knowledge gap: Improving annotation of canine cancer genomic biomarkers through systematic comparative analysis of human cancer mutations in COSMIC.

The accrual of cancer mutation data and related functional and clinical associations have revolutionised human oncology, enabling the advancement of precision medicine and biomarker-guided clinical management. The catalogue of cancer mutations is also growing in canine cancers. However, without direct high-powered functional data in dogs, it remains challenging to interpret and utilise them in research and clinical settings. It is well-recognised that canine and human cancers share genetic, molecular and phenotypic similarities. Therefore, leveraging the massive wealth of human mutation data may help advance canine oncology. Here, we present a structured analysis of sequence conservation and conversion of human mutations to the canine genome through a 'caninisation' process. We applied this analysis to COSMIC, the Catalogue of Somatic Mutations in Cancer, the most prominent human cancer mutation database. For the project's initial phase, we focused on the subset of the COSMIC data corresponding to Cancer Gene Census (CGC) genes. A total of 670 canine orthologs were found for 721 CGC genes. In these genes, 365 K unique mutations across 160 tumour types were converted successfully to canine coordinates. We identified shared putative cancer-driving mutations, including pathogenic and hotspot mutations and mutations bearing similar biomarker associations with diagnostic, prognostic and therapeutic utility. Thus, this structured caninisation of human cancer mutations facilitates the interpretation and annotation of canine mutations and helps bridge the knowledge gap to enable canine precision medicine.

Abstract 3143: A bioinformatic pipeline for identifying change-of-metabolic-function cancer mutations

Abstract Cancer arises through clonal evolution, in some cases selecting change of metabolic function (COMF) mutations that functionally alter proteins to affect cellular metabolism. COMF mutations can impart new catalytic functions that can be applied to enable novel chemical synthesis and metabolic engineering techniques. We previously showed that the oncogenic COMF mutation IDH1 p.R132H could also be used to enable a bio-based method of adipic acid production, a valuable commodity chemical used to synthesize nylon that erstwhile required petroleum-substrates for synthesis. However, there is no high-throughput method to identify COMF mutations. Here, we present METIS (Mutated Enzymes from Tumors In silico Screen), a bioinformatic pipeline to identify COMF mutations from cancer mutational data based on the recurrence rate, genetic conservation, and predicted functionality of the mutations. We applied METIS to 210,354 cancer-derived missense mutations from the COSMIC (Catalogue Of Somatic Mutations In Cancer) database. We identified 4 candidate COMF mutations: the Cbl proto-oncogene E3 ubiquitin ligase (CBL) p.Y371H, Polypeptide N-Acetylgalactosaminyltransferase 17 (GALNT17) p.R228C, solute carrier family 17 member 5 (SLC17A5) p.R364C, and 2-oxoglutarate dehydrogenase-like (OGDHL) p.A400T. To determine these mutations’ effect on the cellular metabolome, we performed unbiased global metabolite profiling using LC-MS/MS and GC/MS of HeLa cells exogenously expressing COMF candidates. OGDHL p.A400T demonstrated significant metabolic changes after FDR correction (q&amp;lt;0.05, two-tailed Welch’s unequal variances t-test with Bonferroni correction). In particular, xanthosine, a key intermediate in purine metabolism commonly used in pharmaceutical development, was increased 2.9-fold (P = 1.2 x 10-9, q = 3.2 x 10-7). Thus, our data suggest OGDHL p.A400T could be used to improve production methods for a useful biochemical. We then deployed METIS2, which featured improved statistical, pathogenicity, and structural analysis tools to all 49 million currently-available cancer mutations within COSMIC, providing a refined panel of six candidate mutations. Consistent with the findings from initial METIS metabolomic screen, METIS2 identified OGDHLp.A400T as a COMF mutation candidate. Overall, our results detail an approach to filter cancer data for mutations that confer metabolic functions, validate that mutations identified in this way can alter the cellular metabolome, and catalog potentially-useful candidate mutations. Moreover, prediction of COMF mutations through METIS can also be applied to elucidate mechanisms of cancer initiation, progression, and/or maintenance to identify potential therapeutic targets. As cancer mutation and structural data continue to accumulate, we expect METIS to increase in its predictive power to accurately find COMF mutations. Citation Format: Kevin J. Tu, Bill H. Diplas, Joshua A. Regal, Matthew S. Waitkus, Christopher J. Pirozzi, Zachary J. Reitman. A bioinformatic pipeline for identifying change-of-metabolic-function cancer mutations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3143.

Abstract 229: Methylation as a surrogate for mutations to identify therapeutic targets

Abstract Background: Patients whose tumors exhibit an inflammatory "hot” phenotype tend to show initial efficacy toward immune checkpoint inhibitors but develop resistance over time, becoming “cold” tumors. Understanding the biology involved in this transition is of interest, especially for the identification of genes where a mutation may be facilitating the transition. As the occurrence of mutations are often rare, identifying these markers in an affected population of patients can be challenging. Here we describe a method using hyper- and hypomethylation, which can mimic the loss and gain of function caused by mutations, as a surrogate to identify mutations potentially involved in the transition of tumors from hot to cold. Methods: Breast cancer methylation and corresponding gene expression data were downloaded from TCGA. Samples were classified as immune hot or cold tumors, the latter being either an immunosuppressive state or immune inert state, by a previously defined gene algorithm [1]. Genes that contributed to immune state classification were compared between gene expression and hypermethylation status by creating a new logistic regression model for the latter. A candidate list of genes that were hypermethylated in cold tumors was created. As a validation step, Cancer Cell Line Encyclopedia (CCLE) expression and mutation data were downloaded from Depmap. Cell lines were classified as hot or cold by the same algorithm and were curated for damaging mutations within the derived gene list. For each candidate gene, the number of damaging mutations in cold cell lines was compared to the number of wild type “hot” cell lines (Chi-squared analysis) and significance was determined by comparing observed values to the distribution of 1000 randomly selected gene sets. Results: Of the final 1562 candidate genes, 420 genes had at least a 5% difference in hot and cold phenotypes between cell lines with and without a damaging mutation in that gene when compared to the methylation data (Z=10.24 , p&amp;lt;0.001). The candidate list of genes was enriched for transcription factors and a previously curated list of tumor suppressor genes. Conclusions: In this proof of principle study, we show that there is a correlation between methylation status in genes that classify tumor immune microenvironment in tumors and the presence of mutations in the same immune effector cell lines which can potentially be used to identify candidate genes involved in the immune status of tumors. While the role of these mutations is unknown, they are enriched for transcription factors and tumor suppressors genes.

A novel cuproptosis-related lncRNAs signature predicts prognostic and immune of bladder urothelial carcinoma.

Bladder Urothelial Carcinoma (BLCA) remains the most common urinary system tumor, and its prognosis is poor. Cuproptosis is a recently discovered novel cell death involved in the development of tumor cells. However, the use of cuproptosis to predict the prognosis and immunity of Bladder Urothelial Carcinoma remains largely unclear, and this study was designed to verify cuproptosis-related long non-coding RNAs (lncRNAs) to estimate the prognosis and immunity of Bladder Urothelial Carcinoma. In our study, we first defined the expression of cuproptosis-related genes (CRGs) in BLCA, and 10 CRGs were up- or downregulated. We then constructed a co-expression network of cuproptosis-related mRNA and long non-coding RNAs using RNA sequence data from The Cancer Genome Atlas Bladder Urothelial Carcinoma (TCGA-BLCA), clinical features and mutation data from BLCA patients to obtain long non-coding RNAs by Pearson analysis. Afterward, univariate and multivariate COX analysis identified 21 long non-coding RNAs as independent prognostic factors and used these long non-coding RNAs to construct a prognostic model. Then, survival analysis, principal component analysis (PCA), immunoassay, and comparison of tumor mutation frequencies were performed to verify the accuracy of the constructed model, and GO and KEGG functional enrichment analysis was used to verify further whether cuproptosis-related long non-coding RNAs were associated with biological pathways. The results showed that the model constructed with cuproptosis-related long non-coding RNAs could effectively evaluate the prognosis of BLCA, and these long non-coding RNAs were involved in numerous biological pathways. Finally, we performed immune infiltration, immune checkpoint and drug sensitivity analyses on four genes (TTN, ARID1A, KDM6A, RB1) that were highly mutated in the high-risk group to evaluate the immune association of risk genes with BLCA. In conclusion, the cuproptosis-related lncRNA markers constructed in this study have evaluation value for prognosis and immunity in BLCA, which can provide a certain reference for the treatment and immunity of BLCA.

O47: Leveraging somatic cancer mutation data to predict the pathogenicity of germline missense variants*

O47: Leveraging somatic cancer mutation data to predict the pathogenicity of germline missense variants*

DNA Repair and Replication-Related Gene Signature Based on Tumor Mutation Burden Reveals Prognostic and Immunotherapy Response in Gastric Cancer.

The genomic variant features (mutations, deletions, structural variants, etc.) within gastric cancer impact its evolution and immunogenicity. The tumor has developed several coping strategies to respond to these changes by DNA repair and replication (DRR). However, the intrinsic relationship between the associated DRR-related genes and gastric cancer progression remained unknown. This study selected DRR-related genes with tumor mutation burden based on the TCGA (The Cancer Genome Atlas) database of gastric cancer transcriptome and mutation data. The prognosis model of seven genes (LAMA2, CREB3L3, SELP, ABCC9, CYP1B1, CDH2, and GAMT) was constructed by a univariate and LASSO regression analysis and divided into high-risk and low-risk groups with the median risk score. Survival analysis showed that overall survival (OS) was lower in the high-risk group than that in the low-risk group. Moreover, patients with gastric cancer in the high-risk group have worse survival in different subgroups, including age, gender, histological grade, and TNM stage. The nomogram that included risk scores for DRR-related genes could accurately foresee OS of patients with gastric cancer. Interestingly, the tumor mutation burden score was higher in the low-risk group than that in the high-risk group, and the risk score for DRR-related genes was negatively correlated with tumor mutation burden in gastric cancer. Next, we further combined the risk score and tumor mutation burden to evaluate the prognosis of gastric cancer patients. The low-risk cohort had a better prognosis than the high-risk cohort in the high tumor mutation burden subgroup. The number of mutation types in the high-risk group was lower than that in the low-risk group. In the immune microenvironment of gastric cancer, more naïve B cells, memory resting CD4+ T cells, Treg cells, monocytes cells, and resting mast cells were infiltrated in the high-risk group. At last, PD-L1 and IAP expressions were negatively correlated with the risk scores; patients with gastric cancer in the low-risk group showed better immunotherapy outcomes than those in the high-risk group. Overall, the DRR-related gene signature based on tumor mutation burden is a novel biomarker for prognostic and immunotherapy response in patients with gastric cancer.

Analyzing TCGA Data to Identify Gene Mutations Linked to Hepatocellular Carcinoma in Asians

Introduction: Liver cancer is the sixth most common and second most fatal type of cancer worldwide. Few treatment options are available as patients with liver cancer are often diagnosed in an advanced stage due to a lack of clinical symptoms. Effectively preventing and treating liver cancer relies heavily on early diagnosis; early diagnosis results from identifying and monitoring high-risk patients. Epigenetic risk factors, such as hepatitis B, hepatitis C, cirrhosis, nonalcoholic fatty liver disease, and alcohol/tobacco abuse, are highly prevalent in Asia and likely cause Asians to have a higher incidence and mortality rate of liver cancer. While these acquired risk factors are relatively well understood, the underlying genetic background of liver cancer in Asians has not been well established or correlated with clinical outcomes. Methods: In this study, we accessed The Cancer Genome Atlas (TCGA) hepatocellular carcinoma clinical and mutation data through TCGAbiolinksGUI. Results: We found that mutations in five genes (TP53, TTN, OBSCN, MUC5B, CSMD1) were statistically linked with increased mortality in Asians compared to non-Asians, four of which (TTN, OBSCN, MUC5B, CSMD1) were also more prevalent in the Asian population. Within the Asian cohort, two gene mutations (TTN, HMCN1) were statistically linked with worse outcomes. We also found that the TP53 mutation predicts worse outcomes within the non-Asian cohort but not within the Asian cohort. Discussion/Conclusion: Our findings can improve cancer care in the Asian population through better disease prognostication, evaluations for potential targeted therapy, and a deeper understanding of liver cancer pathogenesis.

Sequence Neighborhoods Enable Reliable Prediction of Pathogenic Mutations in Cancer Genomes.

Simple SummaryCancer is caused by the accumulation of somatic mutations, some of which are responsible for the disease’s progression (drivers) while others are functionally neutral (passengers). Although several methods have been developed to distinguish between the two classes of mutations, very few have concentrated on using the neighborhood nucleotide sequences as potential discrimination features. In this study, we show that driver mutations’ neighborhood is significantly different from that of passengers. We further develop a novel machine learning tool, NBDriver, which is highly efficient at identifying pathogenic variants from multiple independent test datasets. Efficient and accurate identification of novel pathogenic variants from sequenced cancer genomes would help facilitate more effective therapies tailored to patients’ mutational profiles.Identifying cancer-causing mutations from sequenced cancer genomes hold much promise for targeted therapy and precision medicine. “Driver” mutations are primarily responsible for cancer progression, while “passengers” are functionally neutral. Although several computational approaches have been developed for distinguishing between driver and passenger mutations, very few have concentrated on using the raw nucleotide sequences surrounding a particular mutation as potential features for building predictive models. Using experimentally validated cancer mutation data in this study, we explored various string-based feature representation techniques to incorporate information on the neighborhood bases immediately 5′ and 3′ from each mutated position. Density estimation methods showed significant distributional differences between the neighborhood bases surrounding driver and passenger mutations. Binary classification models derived using repeated cross-validation experiments provided comparable performances across all window sizes. Integrating sequence features derived from raw nucleotide sequences with other genomic, structural, and evolutionary features resulted in the development of a pan-cancer mutation effect prediction tool, NBDriver, which was highly efficient in identifying pathogenic variants from five independent validation datasets. An ensemble predictor obtained by combining the predictions from NBDriver with three other commonly used driver prediction tools (FATHMM (cancer), CONDEL, and MutationTaster) significantly outperformed existing pan-cancer models in prioritizing a literature-curated list of driver and passenger mutations. Using the list of true positive mutation predictions derived from NBDriver, we identified a list of 138 known driver genes with functional evidence from various sources. Overall, our study underscores the efficacy of using raw nucleotide sequences as features to distinguish between driver and passenger mutations from sequenced cancer genomes.

Sequence and Structure-Based Analysis of Specificity Determinants in Eukaryotic Protein Kinases

SummaryProtein kinases lie at the heart of cell-signaling processes and are often mutated in disease. Kinase target recognition at the active site is in part determined by a few amino acids around the phosphoacceptor residue. However, relatively little is known about how most preferences are encoded in the kinase sequence or how these preferences evolved. Here, we used alignment-based approaches to predict 30 specificity-determining residues (SDRs) for 16 preferences. These were studied with structural models and were validated by activity assays of mutant kinases. Cancer mutation data revealed that kinase SDRs are mutated more frequently than catalytic residues. We have observed that, throughout evolution, kinase specificity has been strongly conserved across orthologs but can diverge after gene duplication, as illustrated by the G protein-coupled receptor kinase family. The identified SDRs can be used to predict kinase specificity from sequence and aid in the interpretation of evolutionary or disease-related genomic variants.

HiLDA: a statistical approach to investigate differences in mutational signatures.

We propose a hierarchical latent Dirichlet allocation model (HiLDA) for characterizing somatic mutation data in cancer. The method allows us to infer mutational patterns and their relative frequencies in a set of tumor mutational catalogs and to compare the estimated frequencies between tumor sets. We apply our method to two datasets, one containing somatic mutations in colon cancer by the time of occurrence, before or after tumor initiation, and the second containing somatic mutations in esophageal cancer by sex, age, smoking status, and tumor site. In colon cancer, the relative frequencies of mutational patterns were found significantly associated with the time of occurrence of mutations. In esophageal cancer, the relative frequencies were significantly associated with the tumor site. Our novel method provides higher statistical power for detecting differences in mutational signatures.

Protein interaction disruption in cancer

BackgroundMost methods that integrate network and mutation data to study cancer focus on the effects of genes/proteins, quantifying the effect of mutations or differential expression of a gene and its neighbors, or identifying groups of genes that are significantly up- or down-regulated. However, several mutations are known to disrupt specific protein-protein interactions, and network dynamics are often ignored by such methods. Here we introduce a method that allows for predicting the disruption of specific interactions in cancer patients using somatic mutation data and protein interaction networks.MethodsWe extend standard network smoothing techniques to assign scores to the edges in a protein interaction network in addition to nodes. We use somatic mutations as input to our modified network smoothing method, producing scores that quantify the proximity of each edge to somatic mutations in individual samples.ResultsUsing breast cancer mutation data, we show that predicted edges are significantly associated with patient survival and known ligand binding site mutations. In-silico analysis of protein binding further supports the ability of the method to infer novel disrupted interactions and provides a mechanistic explanation for the impact of mutations on key pathways.ConclusionsOur results show the utility of our method both in identifying disruptions of protein interactions from known ligand binding site mutations, and in selecting novel clinically significant interactions.Supporting website with software and data: https://www.cs.cmu.edu/~mruffalo/mut-edge-disrupt/.

The Cancer Mutation D83V Induces an α-Helix to β-Strand Conformation Switch in MEF2B

MEF2B is a major target of somatic mutations in non-Hodgkin lymphoma. Most of these mutations are non-synonymous substitutions of surface residues in the MADS-box/MEF2 domain. Among them, D83V is the most frequent mutation found in tumor cells. The link between this hotspot mutation and cancer is not well understood. Here we show that the D83V mutation induces a dramatic α-helix to β-strand switch in the MEF2 domain. Located in an α-helix region rich in β-branched residues, the D83V mutation not only removes the extensive helix stabilization interactions but also introduces an additional β-branched residue that further shifts the conformation equilibrium from α-helix to β-strand. Cross-database analyses of cancer mutations and chameleon sequences revealed a number of well-known cancer targets harboring β-strand favoring mutations in chameleon α-helices, suggesting a commonality of such conformational switch in certain cancers and a new factor to consider when stratifying the rapidly expanding cancer mutation data.

Inference of Cancer Progression With Probabilistic Graphical Model From Cross-Sectional Mutation Data

With the advance of high-throughput sequencing technologies, a great amount of somatic mutation data in cancer have been produced, allowing deep analyzing tumor pathogenesis. However, the majority of these data are cross-sectional rather than temporal, and it is difficulty to infer the temporal order of gene mutations from them. In this paper, we first show a probabilistic graphical model (PGM) to infer the temporal order constrains and selectivity relation among the mutation of cancer driver genes which are presented by a directed acyclic graph. We then apply an exponential function based on the mutation probability of these driver genes to obtain their mutation waiting time which can be used to induce mutually exclusive driver pathways. Finally, we evaluate the performance of the PGM both on simulated data and real-cancer somatic mutation data. The experimental results and comparative analysis reveal that the PGM can capture most of the selectivity relation of mutated driver genes which have been validated by previous works. Furthermore, the PGM can provide new insights on simultaneously inferring driver pathways and the temporal order of their mutations from cross-sectional data.

Abstract 2601: COSMIC-3D: exploring cancer mutations in three dimensions for drug design and discovery

Abstract Explosions in the availability of cancer genomic data and protein structure data give us the potential to explore cancer molecular biology at an unprecedented scale, and in atomic detail. We present COSMIC-3D, which combines COSMIC, the most comprehensive cancer mutation database available (http://cancer.sanger.ac.uk), with the wealth of publicly available 3D protein structure data, to create a resource with which the protein-structural nature of cancer can be probed. COSMIC-3D will help us understand which cancer mutations drive the progression of cancers, by identifying which are in functional sites that have an effect on driving cell growth and proliferation, and where they are clustered in protein structures across the proteome, for specific cancer types. Through understanding the effects of cancer mutations on known and predicted drug binding sites, we aim to predict potential new drug targets in cancers, and improve the specificity and efficacy of new or existing drugs, by using protein structure with cancer mutation data to guide mutation-specific drug design. COSMIC-3D is available as a web interface at http://cancer.sanger.ac.uk/cosmic3d, and enables interactive exploration of the cancer mutome in a 3D peptide environment, showing all forms of exonic point mutation. Individual mutant locations can be highlighted as molecular surfaces, while recurrence across nonsynonymous substitutions is visualized as 3D heat-maps. Cancer mutations can also be visualised in combination with precalculated small-molecule “druggable” binding sites, providing a powerful visual approach for development of hypotheses across structural, functional, and drug-resistance impacts of cancer variants. COSMIC-3D shows, for example, the steric occlusion of the binding site of ATP-competitive small-molecule inhibitors in the EGFR kinase domain by the mutation of L858 to arginine; these kinds of insights can be applied to novel cancer targets for any protein-structural mutation of interest. The human cancer structural proteome comprising COSMIC-3D extends to nearly 8,500 human genes; 1/3rd of genes in COSMIC. Over 30,000 human protein structures are available across these genes, to which over 345,000 cancer mutations are mapped. From structure based druggability prediction, over 6,500 proteins in COSMIC are predicted to have small-molecule druggable binding sites, making COSMIC-3D a powerful and exciting resource for the exploration and guidance of drug design and discovery in oncology. Citation Format: Harry C. Jubb, Harpreet Saini, Marcel Verdonk, Simon Forbes. COSMIC-3D: exploring cancer mutations in three dimensions for drug design and discovery [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2601. doi:10.1158/1538-7445.AM2017-2601

Abstract 2436: Exploring genomic alterations in pediatric cancer using ProteinPaint

Abstract Current cancer genome data portals have focused primarily on presenting data generated from adult cancer studies. These portals typically lack features for exploring pathogenic germline mutations, somatic gene fusions, and gene expression profiling, all of which are important biomarkers for risk stratification of pediatric cancer. We have developed ProteinPaint (https://pecan.stjude.org/proteinpaint/), a web service hosting 30,000+ validated somatic SNV/indels and fusion transcripts detected in 1,654 pediatric tumor samples from 17 subtypes, 252 pathogenic or loss-of-function germline lesions detected in &amp;gt;1000 pediatric cancer patients of 21 subtypes, and gene expression profiles derived from RNA-Seq of 928 pediatric tumors. Cancer genomic alterations are shown on novel “disc-on-stem” skewer graphs which were designed to depict the diverse prevalence, complex allelic alteration, and temporal origin of sequence mutations and gene fusions. Adult somatic cancer mutation data from the COSMIC database can be displayed in parallel with pediatric cancer data sets for cross-study comparison. We will demonstrate examples of how ProteinPaint's integrative view of genomic alteration, gene expression and pediatric-adult data comparison has facilitated the evaluation of somatic and germline mutation pathogenicity in a clinical setting. Custom data including sequence mutations in the MAF format used by the Cancer Genome Atlas (TCGA) project, copy number alterations, and structural variations can all be imported and visualized alongside published pediatric and adult cancer data sets. Furthermore, ProteinPaint supports curation and annotation of fusion transcripts predicted from RNASeq data and analysis of tumor clonal evolution with a 2-D plot of mutation frequency of paired diagnosis and relapse samples. ProteinPaint delivers a premium user experience with animation and interactive features for visualizing large cancer mutation datasets, and can serve as a workbench to import, explore and interpret user data. Its framework continues to expand as its intuitive visualization has enabled non-bioinformatics scientists and clinicians to access and manipulate genomic data for discovery and clinical reporting. Citation Format: Xin Zhou, Michael N. Edmonson, Mark R. Wilkinson, Aman Patel, Gang Wu, Yu Liu, Yongjin Li, Zhaojie Zhang, Michael Rusch, Matthew Parker, Jared Becksfort, James R. Downing, Jinghui Zhang. Exploring genomic alterations in pediatric cancer using ProteinPaint. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2436.

Abstract 61: Sequence and structure-guided approach to identify functional mutations in G-protein coupled receptors

Abstract G-protein coupled receptors (GPCRs) account for about 4% of all encoded genes in the human genome with over 800 different types. They activate signal transduction pathways inside the cell that regulate a wide variety of cellular responses and physiological processes. GPCRs are known to play a role in disease progression and are the target of about 40% of drugs on market. However, the implication of GPCRs in tumor initiation and/or progression has not been extensively studied and remains unknown. Due to GPCRs being major drug targets, identifying functional mutations in GPCRs that may lead to tumorigenesis has large therapeutic implications. Nearly 20% of human tumors harbor mutations in GPCRs. However, individual GPCRs may harbor few recurrent mutations making it hard to detect potential functional mutations. Due to this lack of statistical power, it is beneficial to analyze the protein family as a whole and look at recurrent mutations at the same structural site. We can use sequence conservation, protein structure, and cancer mutation data to investigate the mutational landscape of GPCRs and identify conserved mutation hotspots that might be involved in tumorigenesis. We use the positions of structural domains to guide our protein family alignment. Entropy scores, which are used to measure conservation, and somatic mutation densities are computed based on the structure-guided alignment for each position in each domain. Positions that exhibit high somatic mutation density and low entropy are prioritized as possible functional mutations. We analyzed GPCRs from 3,281 tumor samples in over 12 cancer types and have identified three residues located in highly conserved motifs: the DRY, NPxxY, and CWxP motifs that may be involved in the activation of oncogenic pathways. The arginine residue in the DRY motif serves as an “ionic lock”, which keeps GPCRs in an inactive state. Mutations at the arginine could disrupt the inactive conformation leading to a ligand-independent active form and constitutive activation. The NPxxY and CWxP motifs are known to control the equilibrium between the inactive and active states of GPCRs. We plan to further develop our methodology by integrating specific amino acid changes and mutation expression levels during the prioritization process. We will generalize our methodology into a tool called AnaConDAS (Analysis of Conservation in Domain Alignments and Structure) for major druggable protein families. Our lab has created a new tool called HotSpot3D, which clusters mutations and associated drugs based on proximity on a 3D protein structure. We plan to integrate AnaConDAS and HotSpot3D by focusing on the hotpot residues identified by AnaConDAS and analyzing mutations that cluster around them. This approach is novel because it harnesses the statistical power of studying mutations across a protein family and uses 3D protein structure/proximity based analyses to uncover functional mutations in cancer. Citation Format: Sohini Sengupta, Kai Ye, Adam D. Scott, Beifang Niu, Matthew H. Bailey, Michael D. McLellan, Michael C. Wendl, Matthew A. Wyczalkowski, Li Ding. Sequence and structure-guided approach to identify functional mutations in G-protein coupled receptors. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 61. doi:10.1158/1538-7445.AM2015-61

DISEASES: Text mining and data integration of disease–gene associations

Text mining is a flexible technology that can be applied to numerous different tasks in biology and medicine. We present a system for extracting disease–gene associations from biomedical abstracts. The system consists of a highly efficient dictionary-based tagger for named entity recognition of human genes and diseases, which we combine with a scoring scheme that takes into account co-occurrences both within and between sentences. We show that this approach is able to extract half of all manually curated associations with a false positive rate of only 0.16%. Nonetheless, text mining should not stand alone, but be combined with other types of evidence. For this reason, we have developed the DISEASES resource, which integrates the results from text mining with manually curated disease–gene associations, cancer mutation data, and genome-wide association studies from existing databases. The DISEASES resource is accessible through a web interface at http://diseases.jensenlab.org/, where the text-mining software and all associations are also freely available for download.