MapMyCells Gene Mapper

Overview

This code provides a tool to map genes from one species/authority (where by "authority" we mean an institution like ENSEMBL or NCBI) to another species/authority using documented cross-authority and crosss-species (orthologous) gene equvialencies. This code was originally written to allow users to map data collected from an arbitrary species to the species-specific cell type taxonomies supported by the Allen Institute's MapMyCells tool and cell_type_mapper library.

It can, in principle, be applied to other use cases where it would be useful to map genes from one identification scheme to another.

This is a re-implementation in Python of the functionality available through the R GeneOrthology package.

Installation

To install this library, either

A) Clone the repository and run pip install . from the root directory of the repository

or

B) Run

pip install "mmc_gene_mapper @ git+https://github.com/AllenInstitute/mmc_gene_mapper"

To install a specific version, you can run

pip install "mmc_gene_mapper @ git+https://github.com/AllenInstitute/mmc_gene_mapper@{VERSION}"

where {VERSION} is one of the valid tags listed in this repository.

Use

This Jupyter notebook demonstrates how to use the code in this repository.

In broad strokes, you must first create a sqlite database file containing the valid gene mappings as determined from data published by NCBI and ENSEMBL. This can be done either programmatically, according to cell [5] the notebook referenced above, or using the command line tool

python -m mmc_gene_mapper.cli.create_db_file --help

The --help will cause the arguments expected by the command line tool to appear in stdout. A good default would be

python -m mmc_gene_maper.cli.create_db_file \
--db_path data/my_gene_mapper.db \
--local_dir data/downloaded_files/ \
--ensembl_version 114

Before you do this you must download the model that bkbit (a data modelling package written for BICAN) uses to parse ENSEMBL genome annotations. Do this by running the command

bkbit download-ncbi-taxonomy

(bkbit will already be installed as a part of installing this package)

Note: This will produce a 15 GB file on disk.

Once that file has been created, you can instantiate the class MMCGeneMapper (again, see cell [5] of the exampe notebook), and you are ready to map genes.

Level of support

We are providing this tool to the community and any and all who want to use it. Issues and pull requests are welcome, however, this code is also intended as part of the backend for the Allen Institute Brain Knowledge Platform. As such, issues and pull requests may be declined if they interfere with the functionality required to support that service.