search

reimandlab / ActivePathways

Integrative pathway enrichment analysis of multivariate omics data
101 stars 23 forks source link

Problem with gene set thresholds #1

Closed helen-zhu closed 4 years ago

helen-zhu commented 6 years ago

I get the following error ('counts contains negative values. Something went very wrong.') from hypergeometric test whenever I decrease the threshold of the geneset.filter = c(5, 1000) to c(3,1000)

JonBarenboim commented 6 years ago

I'm unable to reproduce the error. Can you provide an example?

reimand0 commented 6 years ago

@helen-zhu can you please provide a reproducible example to @JonBarenboim over email? Thanks!

JonBarenboim commented 6 years ago

Error was caused by duplicated gene names in the scores matrix. Added validation to ensure row names are unique.

helen-zhu commented 6 years ago

Thanks Jon! I forgot - GENCODE has repeated gene names for different ENSEMBL ids. Might be a common problem for people doing pathway analysis directly with experimental results based on these datasets.