Maftools: Efficient analysis, visualization and summarization of MAF files from large-scale cohort based cancer studies

Mutation Annotation Format (MAF) has become a standard file format for storing somatic/germline variants derived from sequencing of large cohort of cancer samples. MAF files contain a list of all variants detected in a sample along with various annotations associated with the putative variant. MAF file forms the basis for many downstream analyses and provides complete landscape of the cohort. Here we introduce maftools an R package that provides rich source of functions for performing various analyses, visualizations and summarization of MAF files. Maftools uses data.table library for faster processing/summarization and ggplot2 for generating rich and publication quality visualizations. Maftools also takes advantages of S4 class system for better data representation, with easy to use and flexible functions. Availability and Implementation: maftools is implemented as an R package available at https://github.com/PoisonAlien/maftools Contact: [email protected] Introduction: With advances in cancer genomics and reduction in cost per base of sequencing technologies, sequencing large cohort of cancer patients has become an efficient way of determining genetic abnormalities associated with the disease [1-4]. Such cohort-based studies often results in large amount of data in the form of somatic/germline variants containing single nucleotide variants (SNP) and small insertion/deletions (INDELS). This data is generally stored in the form of Mutation Annotation Format and provides a complete genomic landscape of the cohort [5]. The Cancer Genome Atlas (TCGA) project has sequenced over 30 different types of cancer and resulting somatic variants are stored as MAF files, with several independent studies following the same [6]. MAF files provides baseline data for many downstream analyses such as driver gene detection, detecting mutually exclusive set of events, mutational signatures and tumor heterogeneity estimation [7-10]. Visualization also plays key role in genomic studies, with researchers often struggling to generate publication quality images, such as oncoplots (also known as waterfall plots), lollipop plots and oncoprints to name a few. As MAF files are getting standardized, current bioinformatic community lacks software to process them. Here, we developed maftools to process, summarize and analyze MAF files, resulting from large cohort based studies. Maftools provides various plotting functions to visualize data stored in MAF files to help researchers generate publication quality images. Functions are also implemented to perform some of the common analyses in cancer studies, including disease associated driver gene detection, mutual exclusivity analysis, and tumor heterogeneity estimation. Along with analysis of MAF files, maftools also provides functions to integrate and visualize of copy number data. Usage of maftools is straightforward with self-explanatory functions and is implemented as an open source R package. . CC-BY-ND 4.0 International license not peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was . http://dx.doi.org/10.1101/052662 doi: bioRxiv preprint first posted online May. 11, 2016;