search

cnio-bu / beyondcell

Beyondcell is a computational methodology for identifying tumour cell subpopulations with distinct drug responses in single-cell RNA-seq and Spatial Transcriptomics data.
Other
40 stars 4 forks source link

Imputing normalized BCS... #67

Closed dys1997 closed 1 year ago

dys1997 commented 1 year ago

When I was using beyondcell to analyze my rds data, I stopped at Imputing normalized BCS... I couldn't continue the analysis, the software did not report an error but kept running (nearly more than ten hours). My rds data is smaller than test data.It only takes about an hour for my computer to analyze the test data。

SGMartin commented 1 year ago

Hello @dys1997

It is a known issue with the current version. Briefly, the KNN imputer is not converging and thus, the imputation never finishes. This is due to too many missing values for either a cell or a signature.

We are aware of this and we'll soon roll out a new version (beyondcell 2.0) adressing this issue among other improvements.

For now, as a workaround, I would:

  1. Try to subset your dataset so that at least one complete case remains for each cell/signature.
  2. If you don't mind, you can share with us your beyondcell object and we'll try to find a sensible cut-off value for the subset.

Can you run complete.cases() over your bc@normalized matrix?

SGMartin commented 1 year ago

Closed due to no further answers from OP