Hidden Markov models lead to higher resolution maps of mutation signature activity in cancer

AMBAR - Interactive Alteration annotations for molecular tumor boards

MOTIVATION

Personalized decision-making for cancer therapy relies on molecular profiling from sequencing data in combination with database evidence and expert knowledge. Molecular tumor boards (MTBs) bring together clinicians and scientists with diverse expertise and are increasingly established in the clinical routine for therapeutic interventions. However, the analysis and documentation of patients data are still time-consuming and difficult to manage for MTBs, especially as few tools are available for the amount of information required.

RESULTS

To overcome these limitations, we developed an interactive web application AMBAR (Alteration annotations for Molecular tumor BoARds), for therapeutic decision-making support in MTBs. AMBAR is an R shiny-based application that allows customization, interactive filtering, visualization, adding expert knowledge, and export to clinical systems of annotated mutations.

AVAILABILITY

AMBAR is dockerized, open source and available at https://sysbio.uni-ulm.de/?Software:Ambar Contact:hans.kestler@uni-ulm.de.

A mutation-level covariate model for mutational signatures

Mutational processes and their exposures in particular genomes are key to our understanding of how these genomes are shaped. However, current analyses assume that these processes are uniformly active across the genome without accounting for potential covariates such as strand or genomic region that could impact such activities. Here we suggest the first mutation-covariate models that explicitly model the effect of different covariates on the exposures of mutational processes. We apply these models to test the impact of replication strand on these processes and compare them to strand-oblivious models across a range of data sets. Our models capture replication strand specificity, point to signatures affected by it, and score better on held-out data compared to standard models that do not account for mutation-level covariate information.

Influence network model uncovers relations between biological processes and mutational signatures

BACKGROUND

There has been a growing appreciation recently that mutagenic processes can be studied through the lenses of mutational signatures, which represent characteristic mutation patterns attributed to individual mutagens. However, the causal links between mutagens and observed mutation patterns as well as other types of interactions between mutagenic processes and molecular pathways are not fully understood, limiting the utility of mutational signatures.

METHODS

To gain insights into these relationships, we developed a network-based method, named GENESIGNET that constructs an influence network among genes and mutational signatures. The approach leverages sparse partial correlation among other statistical techniques to uncover dominant influence relations between the activities of network nodes.

RESULTS

Applying GENESIGNET to cancer data sets, we uncovered important relations between mutational signatures and several cellular processes that can shed light on cancer-related processes. Our results are consistent with previous findings, such as the impact of homologous recombination deficiency on clustered APOBEC mutations in breast cancer. The network identified by GENESIGNET also suggest an interaction between APOBEC hypermutation and activation of regulatory T Cells (Tregs), as well as a relation between APOBEC mutations and changes in DNA conformation. GENESIGNET also exposed a possible link between the SBS8 signature of unknown etiology and the Nucleotide Excision Repair (NER) pathway.

CONCLUSIONS

GENESIGNET provides a new and powerful method to reveal the relation between mutational signatures and gene expression. The GENESIGNET method was implemented in python, and installable package, source codes and the data sets used for and generated during this study are available at the Github site https://github.com/ncbi/GeneSigNet.

A Biterm Topic Model for Sparse Mutation Data

Mutational signature analysis promises to reveal the processes that shape cancer genomes for applications in diagnosis and therapy. However, most current methods are geared toward rich mutation data that has been extracted from whole-genome or whole-exome sequencing. Methods that process sparse mutation data typically found in practice are only in the earliest stages of development. In particular, we previously developed the Mix model that clusters samples to handle data sparsity. However, the Mix model had two hyper-parameters, including the number of signatures and the number of clusters, that were very costly to learn. Therefore, we devised a new method that was several orders-of-magnitude more efficient for handling sparse data, was based on mutation co-occurrences, and imitated word co-occurrence analyses of Twitter texts. We showed that the model produced significantly improved hyper-parameter estimates that led to higher likelihoods of discovering overlooked data and had better correspondence with known signatures.

Regional mutational signature activities in cancer genomes

Cancer genomes harbor a catalog of somatic mutations. The type and genomic context of these mutations depend on their causes and allow their attribution to particular mutational signatures. Previous work has shown that mutational signature activities change over the course of tumor development, but investigations of genomic region variability in mutational signatures have been limited. Here, we expand upon this work by constructing regional profiles of mutational signature activities over 2,203 whole genomes across 25 tumor types, using data aggregated by the Pan-Cancer Analysis of Whole Genomes (PCAWG) consortium. We present GenomeTrackSig as an extension to the TrackSig R package to construct regional signature profiles using optimal segmentation and the expectation-maximization (EM) algorithm. We find that 426 genomes from 20 tumor types display at least one change in mutational signature activities (changepoint), and 306 genomes contain at least one of 54 recurrent changepoints shared by seven or more genomes of the same tumor type. Five recurrent changepoint locations are shared by multiple tumor types. Within these regions, the particular signature changes are often consistent across samples of the same type and some, but not all, are characterized by signatures associated with subclonal expansion. The changepoints we found cannot strictly be explained by gene density, mutation density, or cell-of-origin chromatin state. We hypothesize that they reflect a confluence of factors including evolutionary timing of mutational processes, regional differences in somatic mutation rate, large-scale changes in chromatin state that may be tissue type-specific, and changes in chromatin accessibility during subclonal expansion. These results provide insight into the regional effects of DNA damage and repair processes, and may help us localize genomic and epigenomic changes that occur during cancer development.

A mixture model for signature discovery from sparse mutation data

Mutational signatures are key to understanding the processes that shape cancer genomes, yet their analysis requires relatively rich whole-genome or whole-exome mutation data. Recently, orders-of-magnitude sparser gene-panel-sequencing data have become increasingly available in the clinic. To deal with such sparse data, we suggest a novel mixture model, Mix. In application to simulated and real gene-panel sequences, Mix is shown to outperform current approaches and yield mutational signatures and patient stratifications that are in higher agreement with the literature. We further demonstrate its utility in several clinical settings, successfully predicting therapy benefit and patient groupings from MSK-IMPACT pan-cancer data. Availability: https://github.com/itaysason/Mix-MMM .

Recognizing Pattern and Rule of Mutation Signatures Corresponding to Cancer Types

Cancer has been generally defined as a cluster of systematic malignant pathogenesis involving abnormal cell growth. Genetic mutations derived from environmental factors and inherited genetics trigger the initiation and progression of cancers. Although several well-known factors affect cancer, mutation features and rules that affect cancers are relatively unknown due to limited related studies. In this study, a computational investigation on mutation profiles of cancer samples in 27 types was given. These profiles were first analyzed by the Monte Carlo Feature Selection (MCFS) method. A feature list was thus obtained. Then, the incremental feature selection (IFS) method adopted such list to extract essential mutation features related to 27 cancer types, find out 207 mutation rules and construct efficient classifiers. The top 37 mutation features corresponding to different cancer types were discussed. All the qualitatively analyzed gene mutation features contribute to the distinction of different types of cancers, and most of such mutation rules are supported by recent literature. Therefore, our computational investigation could identify potential biomarkers and prediction rules for cancers in the mutation signature level.

RepairSig: Deconvolution of DNA damage and repair contributions to the mutational landscape of cancer

Cancer genomes accumulate a large number of somatic mutations resulting from a combination of stochastic errors in DNA processing, cancer-related aberrations of the DNA repair machinery, or carcinogenic exposures; each mutagenic process leaves a characteristic mutational signature. A key challenge is understanding the interactions between signatures, particularly as DNA repair deficiencies often modify the effects of other mutagens. Here, we introduce RepairSig, a computational method that explicitly models additive primary mutagenic processes; non-additive secondary processes, which interact with the primary processes; and a mutation opportunity, that is, the distribution of sites across the genome that are vulnerable to damage or preferentially repaired. We demonstrate that RepairSig accurately recapitulates experimentally identified signatures, identifies autonomous signatures of deficient DNA repair processes, and explains mismatch repair deficiency in breast cancer by de novo inference of both primary and secondary signatures from patient data. RepairSig is freely available for download at https://github.com/ncbi/RepairSig.

Mutational Signatures: From Methods to Mechanisms

Mutations are the driving force of evolution, yet they underlie many diseases, in particular, cancer. They are thought to arise from a combination of stochastic errors in DNA processing, naturally occurring DNA damage (e.g., the spontaneous deamination of methylated CpG sites), replication errors, and dysregulation of DNA repair mechanisms. High-throughput sequencing has made it possible to generate large datasets to study mutational processes in health and disease. Since the emergence of the first mutational process studies in 2012, this field is gaining increasing attention and has already accumulated a host of computational approaches and biomedical applications.

Auto-HMM-LMF: feature selection based method for prediction of drug response via autoencoder and hidden Markov model

BACKGROUND

Predicting the response of cancer cell lines to specific drugs is an essential problem in personalized medicine. Since drug response is closely associated with genomic information in cancer cells, some large panels of several hundred human cancer cell lines are organized with genomic and pharmacogenomic data. Although several methods have been developed to predict the drug response, there are many challenges in achieving accurate predictions. This study proposes a novel feature selection-based method, named Auto-HMM-LMF, to predict cell line-drug associations accurately. Because of the vast dimensions of the feature space for predicting the drug response, Auto-HMM-LMF focuses on the feature selection issue for exploiting a subset of inputs with a significant contribution.

RESULTS

This research introduces a novel method for feature selection of mutation data based on signature assignments and hidden Markov models. Also, we use the autoencoder models for feature selection of gene expression and copy number variation data. After selecting features, the logistic matrix factorization model is applied to predict drug response values. Besides, by comparing to one of the most powerful feature selection methods, the ensemble feature selection method (EFS), we showed that the performance of the predictive model based on selected features introduced in this paper is much better for drug response prediction. Two datasets, the Genomics of Drug Sensitivity in Cancer (GDSC) and Cancer Cell Line Encyclopedia (CCLE) are used to indicate the efficiency of the proposed method across unseen patient cell-line. Evaluation of the proposed model showed that Auto-HMM-LMF could improve the accuracy of the results of the state-of-the-art algorithms, and it can find useful features for the logistic matrix factorization method.

CONCLUSIONS

We depicted an application of Auto-HMM-LMF in exploring the new candidate drugs for head and neck cancer that showed the proposed method is useful in drug repositioning and personalized medicine. The source code of Auto-HMM-LMF method is available in https://github.com/emdadi/Auto-HMM-LMF .

Characteristics of mutational signatures of unknown etiology

Although not all somatic mutations are cancer drivers, their mutational signatures, i.e. the patterns of genomic alterations at a genome-wide scale, provide insights into past exposure to mutagens, DNA damage and repair processes. Computational deconvolution of somatic mutation patterns and expert curation pan-cancer studies have identified a number of mutational signatures associated with point mutations, dinucleotide substitutions, insertions and deletions, and rearrangements, and have established etiologies for a subset of these signatures. However, the mechanisms underlying nearly one-third of all mutational signatures are not yet understood. The signatures with established etiology and those with hitherto unknown origin appear to have some differences in strand bias, GC content and nucleotide context diversity. It is possible that some of the hitherto ‘unknown’ signatures predominantly occur outside gene regions. While nucleotide contexts might be adequate to establish etiologies of some mutational signatures, in other cases additional features, such as broader (epi)genomic contexts, including chromatin, replication timing, processivity and local mutational patterns, may help fully understand the underlying DNA damage and repair processes. Nonetheless, remarkable progress in characterization of mutational signatures has provided fundamental insights into the biology of cancer, informed disease etiology and opened up new opportunities for cancer prevention, risk management, and therapeutic decision making.

Network-based approaches elucidate differences within APOBEC and clock-like signatures in breast cancer

BACKGROUND

Studies of cancer mutations have typically focused on identifying cancer driving mutations that confer growth advantage to cancer cells. However, cancer genomes accumulate a large number of passenger somatic mutations resulting from various endogenous and exogenous causes, including normal DNA damage and repair processes or cancer-related aberrations of DNA maintenance machinery as well as mutations triggered by carcinogenic exposures. Different mutagenic processes often produce characteristic mutational patterns called mutational signatures. Identifying mutagenic processes underlying mutational signatures shaping a cancer genome is an important step towards understanding tumorigenesis.

METHODS

To investigate the genetic aberrations associated with mutational signatures, we took a network-based approach considering mutational signatures as cancer phenotypes. Specifically, our analysis aims to answer the following two complementary questions: (i) what are functional pathways whose gene expression activities correlate with the strengths of mutational signatures, and (ii) are there pathways whose genetic alterations might have led to specific mutational signatures? To identify mutated pathways, we adopted a recently developed optimization method based on integer linear programming.

RESULTS

Analyzing a breast cancer dataset, we identified pathways associated with mutational signatures on both expression and mutation levels. Our analysis captured important differences in the etiology of the APOBEC-related signatures and the two clock-like signatures. In particular, it revealed that clustered and dispersed APOBEC mutations may be caused by different mutagenic processes. In addition, our analysis elucidated differences between two age-related signatures-one of the signatures is correlated with the expression of cell cycle genes while the other has no such correlation but shows patterns consistent with the exposure to environmental/external processes.

CONCLUSIONS

This work investigated, for the first time, a network-level association of mutational signatures and dysregulated pathways. The identified pathways and subnetworks provide novel insights into mutagenic processes that the cancer genomes might have undergone and important clues for developing personalized drug therapies.

A Sticky Multinomial Mixture Model of Strand-Coordinated Mutational Processes in Cancer

The characterization of mutational processes in terms of their signatures of activity relies mostly on the assumption that mutations in a given cancer genome are independent of one another. Recently, it was discovered that certain segments of mutations, termed processive groups, occur on the same DNA strand and are generated by a single process or signature. Here we provide a first probabilistic model of mutational signatures that accounts for their observed stickiness and strand coordination. The model conditions on the observed strand for each mutation and allows the same signature to generate a run of mutations. It can both use known signatures or learn new ones. We show that this model provides a more accurate description of the properties of mutagenic processes than independent-mutation achieving substantially higher likelihood on held-out data. We apply this model to characterize the processivity of mutagenic processes across multiple types of cancer.

DNA Repair Footprint Uncovers Contribution of DNA Repair Mechanism to Mutational Signatures

Cancer genomes accumulate a large number of somatic mutations resulting from imperfection of DNA processing during normal cell cycle as well as from carcinogenic exposures or cancer related aberrations of DNA maintenance machinery. These processes often lead to distinctive patterns of mutations, called mutational signatures. Several computational methods have been developed to uncover such signatures from catalogs of somatic mutations. However, cancer mutational signatures are the end-effect of several interplaying factors including carcinogenic exposures and potential deficiencies of the DNA repair mechanism. To fully understand the nature of each signature, it is important to disambiguate the atomic components that contribute to the final signature. Here, we introduce a new descriptor of mutational signatures, DNA Repair FootPrint (RePrint), and show that it can capture common properties of deficiencies in repair mechanisms contributing to diverse signatures. We validate the method with published mutational signatures from cell lines targeted with CRISPR-Cas9-based knockouts of DNA repair genes.