Question about assigning gene names, tximport and DESeq2

Question

[bioinfo]

Hello,

I am using kallisto to align my samples and then I use tximport to assign gene names. To do this I use biomart to use the appropriate ensembl version. However, after ensembl version 103 I have been getting a row at the beginning of the output file that does not have a row name (no gene name) which after searching I understood that these are the counts from the genes that do not have an external_gene_name associated with the transcript ID. I also use tximport and biomart when improrting files into DESeq. I do see that the empty row does come up in the results. For example, when I save the results I will have a row with no gene name but it will have the statistics from DESeq2. Is that a problem? Does it need to be removed?

Thank you

[Istvan Albert]

we can't tell what is wrong without seeing the code, but it makes little sense to have an empty gene name row,

for genes without a external name one ought to use some sort of substitute name (the original gene id, for example)

[bioinfo]

When i create the t2g file to assign the gene names I do this because I use ensembl version 112:

mart <- biomaRt::useMart("ensembl", host= "https://may2024.archive.ensembl.org", "hsapiens_gene_ensembl")
t2g <- biomaRt::getBM(attributes = c("ensembl_transcript_id", "external_gene_name"), mart = mart)
t2g <- dplyr::rename(t2g, target_id = "ensembl_transcript_id", ext_gene = "external_gene_name")

The t2g file that is created contains transcripts that do not have external gene names.

When I was using ensembl version 101 there were not transcripts with empty gene names so I did not have that issue.