geshtin
commented
3 years ago ls, fyi. I have been digging into it a bit more and came to some interesting observations. I split up my file into smaller parts to see if somehwere there are some "strange entries" Sometimes I came accross non-unique gene names (that is: two different gene_id entries, that have the same gene_name), But this was not a problem with just a few instances (when there were only, let's say, about 50 instance with duplicate names the refflat was created). But the entire file (more than 250 duplicate names) it crashes. Removing all these duplicate names (replacing all the gene_names by the gene_id entry): running fine. I also noticed that "half entries" so starting at exon as first line (and no corresponding initial gene line): no problem. So setting the VALIDATION_STRINGENCY=STRICT to some other value might also work (I did not test this), What is the limit of these type or errors?? Raymond
seb-mueller
I have problems creating the refflat from the gtf file. It has been reported more often, I followed all suggestion, and as far as I can see the gtfs's look OK. I have one where it works:
with the following error:
Both gf's have the same fields in column 9. I see: "GTFParser Seen many non-increasing record positions" what does this exactly mean? So, what is wrong here? Is the order of the gene_id, transcript_id, gene_name, transcript_name important. I assume the gene lines do not need a transcript id (btw adding this information to the gene lines did not solve my problems). For transcript_name I simply used transcript_id Also the one entry where there is a ";" in the gene name was removed. And finally, there was a mismatch in the fasta and the gtf wrt the scaffold names (for Mt and Pt) this was also corrected.
Any suggestion is greatly appreciated, Raymond