CIViCutils: Matching and downstream processing of clinical annotations from CIViC

Abstract

Background: With the advent of next-generation sequencing, profiling the genetic landscape of tumors entered clinical diagnostics, bringing the resolution of precision oncology to unprecedented levels. However, the wealth of information generated in a sequencing experiment can be difficult to manage, especially if hundreds of mutations need to be interpreted in a clinical context. Dedicated methods and databases are required that assist in interpreting the importance of a mutation for disease progression, prognosis, and with respect to therapy. Here, the CIViC knowledgebase is a valuable curated resource, however, utilizing CIViC in an efficient way for querying a large number of mutations needs sophisticated downstream methods. Methods: To this end, we have developed CIViCutils, a Python package to query, annotate, prioritize, and summarize information from the CIViC database. Our package provides functionality for performing high-throughput searches in CIViC, automatically matching clinical evidence to input variants, evaluating the accuracy of the extracted variant matches, fully exploiting the available disease-specific information according to cancer types of interest, and in-silico predicting drug-target interactions tailored to individual patients. Results: CIViCutils allows the simultaneous query of hundreds of mutations and is able to harmonize input across different nomenclatures. Moreover, it supports gene expression data, single nucleotide mutations, as well as copy number alterations as input. We utilized CIViCutils in a study on the bladder cancer cohort from The Cancer Genome Atlas (TCGA-BLCA), where it helped to extract clinically relevant mutations for personalized therapy recommendation. Conclusions: CIViCutils is an easy-to-use Python package that can be integrated into workflows for profiling the genetic landscape of tumor samples. It streamlines interpreting large numbers of variants with retrieving and processing curated CIViC information.