Identifier Refinery Conversion Matrixes

<strong>identifier-refinery</strong> Tools and assets for easy and reproducable gene identifier conversion. <strong>Methods</strong> This repository is used to build matricies which can convert between different gene identifiers. These conversion matricies are built by: Randomly choosing raw CEL files from NCBI GEO for a given platform accession code (in <code>/cels</code>) Reading the CEL header and joining Brainarray (e.g., <code>hgu133plus2hsensgprobe</code>) and Bioconductor (e.g., <code>hgu133plus2.db</code>) (x, y) coordinates Finding intersecting probe identifiers Extracting supported identifiers and probe IDs from the Bioconductor package Filtering on probe IDs and Ensembl Gene IDs in Brainarray Writing the output to a conversion TSV file Check that all output conversion TSV files have a shared SHA1 <strong>Repository Contents</strong> <strong>Source Files</strong> The <code>cels</code> directory contains raw CEL files taken from GEO. The list of supported platforms is in <code>supported_microarray_platforms.csv</code>. Source files can be acquired by running the <code>acquire_cels.py</code> script. <strong>Docker Image</strong> The conversion scripts are run on custom Docker images. Two Dockerfiles are provided in this repository - <code>base</code> Docker image, which is used to install the quire R dependancies, and the <code>pd</code> image, which is used to build the required databases for a given platform. <strong>Conversion Scripts</strong> A <code>build_and_convert.py</code> script is provided, which build a unique Docker image for each package, mount the downloaded CEL files as a volume, and then run the gene conversion script <code>R/gene_convert.R</code> inside the image and output the master conversion matrix. Output TSV files live in <code>cels/out/</code>. <strong>Reproducing</strong> The entire process can be reproduced by running the following command script from a fresh checkout of this repository. It will take some time: <pre><code>$ ./generate_matricies_from_scratch.sh </code></pre> You can also choose to only build a specific platform, ex.,: <pre><code>$ ./generate_matricies_from_scratch.sh celegans </code></pre> <strong>Identifiers</strong> Released assets in this repository are availble under the DOI, <code>xyz:1.2.3.4</code>, which can be seen on Zenodo here. <strong>Related Projects</strong> AlexsLemonade/refinebio <strong>Copyright</strong> <code>identifier-refinery</code> output assets are released under a CC0 1.0 Universal license. All code is released under the BSD 3-clause license. Input assets are property of the original providers to NCBI GEO, but may be freely downloaded and redistributed unless otherwise noted. https://github.com/AlexsLemonade/identifier-refinery