dhimmel / entrez-gene

Processing the human Entrez Gene subset

Creating user-friendly Entrez Gene datasets for humans

Entrez Gene is the NCBI database of gene-specific information. It provides "tracked, unique identifiers for genes" and reports "information associated with those identifiers for unrestricted public use [source]." We use Entrez Gene as the primary gene vocabulary for our drug repuposing research.

This repository creates user-friendly datasets from Entrez Gene. We currently focus on human genes only.

The python notebook process.ipynb executes the analysis. Files downloaded from external locations are stored in download. The following created datasets reside in data:

• genes-human.tsv: human genes with a select set of fields storing additional attributes
• symbols-human.tsv: a table of GeneID, symbol, and symbol type (synonym or primary)
• symbols-human.json: a Symbol–GeneID mapping of primary symbols only
• synonyms-human.json: a Symbol–GeneIDs mapping for synonyms
• symbol-map.json: a Symbol–GeneID mapping with approved symbols and unambiguous synonyms
• xrefs-human.tsv: mappings to external resources

entrez-gene's Issues

How did you decide which data from the Entrez Gene site to include within hetio?

I am currently scripting auto-update scripts for hetio and I am currently working on the Gene nodes, So, I was looking at the process.ipynb file, and I was wondering how exactly you decided which of the 56532 gene entries from the NCBI website got input into hetio. I realized that you got rid of any gene that was not a Homo sapiens gene and I am guessing you got rid of any gene that is not "protein-coding." However, I did notice a few genes within the SPOKE network (for instance: LOC102725191) that are not protein-coding genes, and I was wondering how exactly you decided which ones went into the network from the Entrez Gene database.