Hi all, Do you know if there is a file available for download in GTEX that answers the question:
the gene/transcript G1 is most probably expressed (mRNA) in tissue T1
Or, if not, what would be the best way to build that resource ? Something like;
SCN5A; ENSG00000183873 ; http://purl.obolibrary.org/obo/BTO_0000562; heart; http://snomed.info/id/410515003;Known_present
Thanks ! P
•
written
by
Pierre Lindenbaum
The Human Protein Atlas (link to the download page) is possibly the most comprehensive aggregator at the tissue level.
With resolution at the single-cell level, I would dare you that your question is ill-posed or at least less interesting: single cell atlas. There could be cell types in a tissue that express a gene while others don't.
The more relevant question is framed as which gene set is a good marker for a cell-type, irrespective of exact level, see https://www.10xgenomics.com/analysis-guides/web-resources-for-cell-type-annotation
Thanks michael, "https://www.proteinatlas.org/download/tsv/normal_ihc_data.tsv.zip" , looks more or less what I needed.
Gene read counts are available for tissues (as separate files) on this page (expand gene read counts section) : https://gtexportal.org/home/downloads/adult-gtex/bulk_tissue_expression
Download and parse the files to create a matrix?
GenoMax thanks. After all those years I'm not familiar with expression data. Which file do you think I should download and how can I set a treshold to say : this gene/protein is probably (not) expressed in a specific tissue.
That is a tricky question. Absence of counts for a gene does not absolutely mean that the gene is not expressed. It is not detectable at the sampling level used.
Using TPM values may make more sense if you want to compare across tissues (TPM are also available) and say something about expression.
Log in to answer this question.
Bgee provides calls of active expression (presence / absence) over many curated libraries (including those which are in the HPA or most GTEx, recurated). For each call we also provide a p-value, i.e. the library-specific probability of having this TPM with no active expression. We also provide the calls and p-values aggregated over libraries, to provide you a unique answer per condition, e.g. in a given organ, tissue, cell type, etc.
You can find these calls here: https://www.bgee.org/search/raw-data?pageType=proc_expr_values as well as on our FTP https://www.bgee.org/ftp/current/download/processed_expr_values/ or through our R package https://bioconductor.org/packages/BgeeDB/
For GTEx data specifically, you can find expression values by tissue here: https://www.bgee.org/search/raw-data?pageType=proc_expr_values&data_type=RNA_SEQ&data=e13a295b0ee7d231ef716d71d1aa984d6f920001