salmonIndex seq name from genes_bed2fasta

[hanrong498]

Hi!

Sorry for posting many issues in the past few days :) I had an error from salmon index:

[2024-10-29 13:22:09.436] [puff::index::jointLog] [info] Running fixFasta
[2024-10-29 13:22:09.444] [puff::index::jointLog] [error] In FixFasta, two references with the same name but different sequences: RefSeq. We require that all input records have a unique name up to the first whitespace (or user-provided separator) character.

The problem turns out to in genes.fa, as they have the same sequence names:

head /scratch/hhu/Xenopus_laevis_v10_1_lambda_spikein/annotation/genes.original.fa
>RefSeq
acaaactacagctcccagcaaccCTTTGCCACCTCGATAGCAAGAAATGTAACAGTTCTTTCAGTGCAACTGAACTCCAAGCTATTAAACTAG
>RefSeq
TTGAGCCACCCACATCATGGACTTTGCCCCTGAGGGCAGATCAGACCCGACAGAGGGCTTATGGGTTAAATAAATCACCTATTGCactaaa
..

I think the command in the genes_bed2fasta:

bedtools getfasta -name -s -split -fi /scratch/hhu/Xenopus_laevis_v10_1_lambda_spikein/genome_fasta/genome.fa -bed <(cat /scratch/hhu/Xenopus_laevis_v10_1_lambda_spikein/annotation/genes.bed | cut -f1-12) | sed 's/(.*)//g' | sed 's/:.*//g' > annotation/genes.fa 2> annotation/logs/bed2fasta.log

It cannot deal with genes of name like this:

Chr4L   15610   37088   RefSeq:XR_005966836.1   .       -       15610   37088   255,0,0 4       144,83,50,796   0,446,2433,20683
Chr4L   40727   57680   RefSeq:XM_041589621.1   .       +       40727   57680   255,0,0 9       283,98,133,116,111,101,102,139,278      0,1109,6513,9213,9409,11052,12648,16142,16676

Maybe this could be a potential problem for others using not so common GTF of other species..

Thanks a lot!
Hanrong

[katsikora]

Hi Hanrong,

in this case, I'd recommend deleting the RefSeq prefix from contig names, such that the resulting contig names are unique.

Best wishes,

Katarzyna