EV Marker Prediction -- Supplemental Files

This repository contains required reference files needed for the EV marker prediction pipeline at https://github.com/klfarenhem/EV_marker_prediction In the EV_marker_prediction directory, run: wget "https://zenodo.org/api/records/21498982/files-archive" -O repository.zip mkdir ref unzip repository.zip -d ref/ rm repository.zip unzip ref/SCimilarity_atlas_subsample.zip -d SCimilarity/SCimilarity_atlas_subsample/ rm ref/SCimilarity_atlas_subsample.zip This repository includes: A subsample of the full SCimilarity Human Tissue Atlases, which are available in full at https://zenodo.org/records/10895214. However, running the pipeline from such large datasets is very memory-intensive, and it is generally recommended to use the included subsamples. The subsamples were calculated using ./supp_code/SCimilarity_atlas_subsample.py. Precalculated tau values from GTEx. Raw data was downloaded from https://www.gtexportal.org/home/downloads/adult-gtex%23qtl("Median gene-level TPM by tissue") and processed into tau values using ./supp_code/GTEx_tau_calc.ipynb Tissue specificity (TS) proteomics and RNAseq scores from Snyder (2020), downloaded directly from https://www.cell.com/cell/fulltext/S0092-8674(20)31078-3. This is labeled table S4, though the file download is called mmc5.xlsx. The file was only processed to remove the rows above the header in Sheet A and save as a csv. Gene annotations for cross-reference, including whether the protein is annotated as transmembrane. This TM annotation was added in ./supp_code/260428_annotation_update.ipynb. Briefly, proteins are annotated as transmembrane if they have (by SwissProt/UniProt Reviewed annotation) at least one transmembrane domain, at least one extracellular region, and have at least one isoform whose subcellular localization includes "cell membrane" but is not secreted. Isoform information is included in the output. A list of proteins detected in EV samples, from !!!!!!!!!!!!