search

Danko-Lab / BayesPrism

A Fully Bayesian Inference of Tumor Microenvironment composition and gene expression
147 stars 46 forks source link

gene names of reference and mixture do not match! #87

Open rmontagn opened 4 months ago

rmontagn commented 4 months ago

Hello,

First I would like to thank you for developping BayesPrism v2.2. It allows me to use large atlases as reference for deconvolution.

However, when I tried to use it, I got an error for some deconvolution tasks:

Error in new.prism(reference = dfRef, mixture = dfBulk, input.type = "count.matrix",  : 
  Error: gene names of reference and mixture do not match!

The error arose when I tried to deconvolve 2 bulk RNA-Seq data and 2 pseudo-bulk RNA-Seq datasets generated by aggregating the reference single cell signal.

While the deconvolution algorithm ran for the bulk RNA-Seq arrays, I can't deconvolve the pseudo-bulk data.

My bulk and reference arguments are dataframes with exactly the same genes as columns and in the same order. So I do not understand what could go wrong.

Do you have any idea ?

Thank a lot

tinyi commented 4 months ago

Dear user,

Thank you for your interest in our method.

Could you kindly dump your input into an rdata file and share it with me (tc532@cornell.edu) for troubleshooting?

Best,

Tinyi