Formatting of fasta file for library-free search/library generation

[SebastianHoernstein]

Hi,

I am running DIA-NN for the first time, So far it worked and I got output. However it seems that the program has difficulties to parse the description lines correctly from the fasta. In some cases there are no description lines, should I introduce an arbitrary? Further the desc lines which are present are split into distinct fractions. All this causes problems in the tsv output.
Is there any recommendation of a format regarding separation of gene identifier and description line?
Thanks in advance.

[vdemichev]

Hi Sebastian,

It works with UniProt format. For all other formats it will correctly parse sequence IDs, and then you can augment DIA-NN report with descriptions, etc, just by loading the FASTA file in question using some R or Python package and matching sequence IDs reported by DIA-NN to those in the loaded FASTA.

Best,
Vadim

[SebastianHoernstein]

Hi,

thanks for the input. I constructed a fake Uniprot fasta using a PERL skript, it works now in a way that I can use it. Sorry but I would have another question, my Database contains 49.589 entries, but when I run the spectral library generation it would contain only 49.377 proteins, how to explain this discrepancy and how could I check which poteins are "missing"?

Best and thanks again

[vdemichev]

Probably no peptides satisfying the digestion requirements originating from some proteins. You can export .predicted.speclib to .tsv to see what's actually being generated (might be a very large .tsv file though)

[SebastianHoernstein]

ok thanks, will try this.

best

Sebastian