Most genes with entrez id in KEGG database don't have a pathway assigned. Is there a tool already available that uses textmining to assign pathway to genes.

Question

[Saad Khan]

Hi,

I was working with bioconductor package KEGGREST in trying to find gene 2 pathway mappings. In Kegg according to KEGGREST there are only 7364 genes which are present in the 317 pathways. There are a total of 39556 genes in KEGG I was wondering how can I find pathways for the rest of the genes or if there is a latest method that I can use to predict pathway assignment for a gene.

thanks!

[theobroma22]

I’m not sure of any method to do that. I do find genes that are not assigned, or curated, to pathways are mostly oxidation-reduction related genes. You could search NCBI or Uniprot to see if there is any curation outside of KEGG. Alternatively, a gene enrichment analysis might be helpful, but again gene assignment to pathways is based on empirical data so it could be that those genes haven’t been shown to participate in any pathway. I do find most of the assigned genes are enzymes and not proteins. What type of genes are you looking to assign to pathways?

[Saad Khan]

We did contrast mining between cases and controls for a neurological disorder in which we found SNP associations between a list of gene pairs. We want to do module enrichment for those gene pairs. Many of the genes identified in these gene pairs do not have a reactome_id or kegg pathway id associated with them as per the results from KEGGREST and Reactome database so we want to putatively assign pathways to those genes based on text mining etc I was wondering if there is a method that already does that. Since it seems to be such a common issue. Thanks!

[natasha.sernova]

See this post, it may help

pathway mapping using KEGG