Error with --search option about dysgu HOT 11 CLOSED

@kcleal I got it working after I downloaded the reference, uncompressed it, compressed it with bgzip and indexed it with samtools.

from dysgu.

Hi @husamia, does the reference genome use 17 instead of chr17? The bam file looks like it uses the chr17 convention, but not the reference? check with head Homo_sapiens_assembly19_1000genomes_decoy.fasta.fai

from dysgu.

@kcleal I've tried replacing chr17 and just 17 in the bed file and it didn't help. I even used samtools view -b chr17 original.bam > chr17.bam and removed the --search and still getting the same error!

from dysgu.

I think the bam file is consistent with chr17 - the rest of the pipeline works ok. I was wondering about the reference genome. The error message is saying that chr17 is not in the fasta file. Tne functionref_genome.get_reference_length is throwing the key error. Could you confirm this?

from dysgu.

@kcleal here is the check

head /mnt/d/Research/Homo_sapiens_assembly19_1000genomes_decoy.fasta.fai
1 249250621 52 100 101
2 243199373 251743232 100 101
3 198022430 497374651 100 101
4 191154276 697377358 100 101
5 180915260 890443229 100 101
6 171115067 1073167694 100 101
7 159138663 1245993964 100 101
8 146364022 1406724066 100 101
9 141213431 1554551781 100 101
10 135534747 1697177401 100 101

from dysgu.

Ah ok, thats the problem - the 'chr' is missing from the chromosome names. You can add this in using sed/awk, and then re-index the genome. It should read chr1 249250621 52 100 101. Alternatively, you should probably use the same reference genome that the sample was aligned to. You can normally check using samtools view -H and look at the command used during mapping along with the reference genome.

from dysgu.

@kcleal the reference used is the generic hg19.fasta which I have two copies of downloaded from different sources and both are the same. furthermore, I can open the file in IGV and the chr17 file created with samtools in IGV as well

from dysgu.

I think IGV does automatic conversion between the two representations. This is not supported in dysgu unfortunately

from dysgu.

@kcleal this is an issue.

the reference is generic named hg19.fasta which is not provided with the data. I used the reference with many other tools just fine. so can a feature added to do the conversion. Provide me me the awk/sed command to add the chr to the file?

from dysgu.

Its a problem I have encountered before with other genomics analysis, the different representations can be a pain. However doing automatic conversion can also cause problems. I recommend
you download the hg19 with the chr representation from ucsc TableBrowser.

from dysgu.

Actually from here: http://hgdownload.cse.ucsc.edu/goldenpath/hg19/bigZips/

from dysgu.