Closed LacquerHed closed 1 year ago
I think the issue lies with the alternative format of GTF. I would need to update the function to accommodate the alternative format of GTF. Please may you share your MARVEL object after the running the CreateMarvelObject.10x function please? Here or my email (sean.wen@astrazeneca.com) will be fine.
Hi Sean, thanks for the reply! I am emailing you a google link for the object from the tutorial now, however its quite large at 7.8gb. So if the GTF format has changed, isn't your tutorial data using the older GTF that worked when you developed the tutorial? Thanks again!
Actually I believe I just adjusted the function to account for GTF:
Previously: strsplit(rownames(df.sj.count), split = ":", fixed = TRUE)
Changed to: strsplit(as.character(df.sj.count), split = ";", fixed = TRUE)
Seems to be running now, does this seem like the correct adjustment?
Here is a snipped of the GTF:
1 gene_id "ENSG00000243485"; gene_version "5"; gene_name "MIR1302-2HG"; gene_source "havana"
Update: It does run but does not finish with this modification, so not sure its correct. Seems to run out of memory and shutoff.
I see that the splice junction coordinates are missing from the rownames of the splice junction count matrix.
am wondering if the splice junction coordinate file (featureData.txt
) was formatted as per the tutorial (see "Splice junction counts" section), and was subsequently specified with the sj.count.feature
option of the CreateMarvelObject.10x
function (see "Create MARVEL object" section)?
Hi getting the following error when running this portion of tutorial of test data:
marvel <- AnnotateSJ.10x(MarvelObject=marvel)
Creating splice junction metadata... Error in strsplit(rownames(df.sj.count), split = ":", fixed = TRUE) : non-character argument
Thanks!