humanlongevity / HLA

xHLA: Fast and accurate HLA typing from short read sequence data
Other
103 stars 52 forks source link

run get-reads-alt-unmap.sh in container #35

Closed tmhmxg59 closed 5 years ago

tmhmxg59 commented 6 years ago

I'm struggling to get typing information from hg19 aligned bams. I think I'd like to run get-reads-alt-unmap.sh in the docker container - I don't have all the dependencies on my machine.

I think I need to change the entrypoint, but I'm not sure the syntax. The coordinates in the .sh script appear to cover all genes in both references.

I tried several variations of: $docker run -itv pwd:pwd pwd -w humanlongevity/hla --entrypoint /bin/bash ls docker: invalid reference format: repository name must be lowercase. See 'docker run --help'.

But doesn't seem like I've successfully changed the entrypoint.

adeschen commented 6 years ago

Hi,

xHLA only works with hg38 reference.

The get-reads-alt-unmap.sh script needs to be used when you have a BAM aligned on hg38 reference with alt contigs.

Best, Astrid

tmhmxg59 commented 6 years ago

Conversion from 37 to 38 worked for me for one bam. I used: docker run -v ~/bams:/opt/bin/myData -it --entrypoint=/bin/bash humanlongevity/hla to open a container with my local directory /bams bound to /opt/bin/myData.

I did have to remove the memory detection part of the script as it gave an error at line 17. And add the fna as @cjieming noted.

On a second data set, though, I'm loosing about half of the reads somewhere in the script. My first go I added a few reads as expected. The only thing I can think of is: when I ran the first time I was using a 37 bam PrintRead to chr6. The first step in the script is to trim, though, so I'm not sure what the issue is.

tmhmxg59 commented 5 years ago

Second data set had chromosome names as number only, so modified get-reads-alt-unmap.sh to 6 instead of chr6.