cellme

Convert a human cell line identifier into a truth-track VCF of its known mutations.

Introduction

cellme builds a truth/known VCF for a cell line by resolving a name such as MOLT-4 to a Cancer Cell Line Encyclopedia sample, fetching that sample's somatic mutations from the cBioPortal REST API, and writing them as a sorted VCF. The mutation source is the ccle_broad_2019 study. CCLE reports its coordinates against hg19 (GRCh37); by default cellme lifts them to hg38 (GRCh38), and --build hg19 keeps the native coordinates.

Installation

Install the release from PyPI with pip:

pip install cellme

Usage

Write an hg38 truth track for the cell line MOLT-4 to a file:

cellme "MOLT-4" --build hg38 --reference /ref/hg38.fa --output MOLT-4.hg38.vcf.gz

The build keys are hg38 and hg19 (the Ensembl names GRCh38 and GRCh37 are accepted as aliases). The query is matched case-insensitively on the leading cell line token, so MOLT-4, MOLT4, and the full MOLT4_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE sample ID all resolve to the same sample.

VCF INFO Fields

Every record is annotated with compact, self-describing INFO fields.

The ID column is set to a stable gene:proteinChange token where one is available, for example TP53:R306*, and falls back to gene:chrom:pos otherwise.

Development and Testing

See the contributing guide for more information.