iCOMIC

iCOMIC implements reproducible workflows for processing whole-genome and transcriptome sequencing (FASTQ) to perform alignment, variant calling, transcript quantification, and driver/passenger mutation prediction for cancer genomics research.

Key Features:

• Snakemake integration: Core workflows for whole-genome and transcriptome analyses are embedded within the Snakemake workflow management system.
• Whole-genome and transcriptome support: Provides end-to-end processing for both DNA-Seq and RNA-Seq data starting from FASTQ files.
• Modularity and customization: Pipeline modules can be replaced or modified to accommodate specific analysis requirements.
• Machine learning integration: Incorporates cTaG and NBDriver to predict driver and passenger mutations using mutational context and data from tumor suppressor genes and oncogenes.
• Support for standard tools: Implements established tools including BWA MEM, GATK HaplotypeCaller, HISAT2, STAR, StringTie, and ballgown for alignment, variant calling, and transcript quantification.

Scientific Applications:

• Cancer genomics: Identification and prioritization of driver and passenger mutations in oncogenes and tumor suppressor genes.
• Variant discovery and benchmarking: Accurate SNP and indel calling for whole-genome analyses and performance assessment against benchmark datasets.
• Transcriptomics and expression analysis: Quantification and downstream analysis of RNA-Seq data using HISAT2/STAR with StringTie and ballgown workflows.

Methodology:

Workflows are managed with Snakemake; DNA-Seq uses a BWA MEM → GATK HaplotypeCaller pipeline; RNA-Seq uses HISAT2–StringTie–ballgown and STAR–StringTie–ballgown pipelines; machine learning methods cTaG and NBDriver predict driver/passenger mutations from mutational context and oncogene/tumor suppressor data; benchmarking reported F1 scores of 0.971 for indels and 0.988 for SNPs on the GIAB NA12878 dataset and an RNA-Seq correlation r=0.85 on the human monocyte dataset SRP082682.