findPath using fastqs in sub-directories

Question

[kathrynm]

Hello,

I am using the Docker container for PHG and was wondering if there was a way to mount a single parent directory of fastq files, where the fastqs are organized into sub-directories based on what plate they came from. I tried adding the sub-directory information to the key file in front of the file name, but PHG gave an error about not having directory information in the file name.

For the time being, I will probably just move all of my fastqs into a single folder, but would be interested in a method for referencing multiple directories of fastqs via a single mount point of the parent folder in the future.

Thanks, Kathryn

Answer 1

[pjb39]

PHG only searches a single directory for fastq files. However, extending that to a recursive search seems like a good idea. Also, implementing that would not be difficult. I will add that to the PHG task list. Oops - the code already does a recursive search. See comment below.

[pjb39]

I answered too quickly. I reviewed the PHG code and found that it is already doing a recursive search. You should be able to specify the top level directory as the fastqDir for path finding and as long as the key files are in that directory or a subdirectory they should be found. You are correct that the subdirectory names should not be included in the key file.

[kathrynm]

Thanks for the quick reply Peter.

I'm having trouble somewhere- my key file lives in the top level directory and the file names don't have the path attached. I'm getting an error: ERROR: failed to open file '/phg/inputDir/imputation/fastq/11367.8.203207.ATGCACG-TCGTGCA.r1_cutadapt.fq': No such file or directory ERROR: failed to map the query file

Thoughts?

Edit- My docker image is from early September if that matters.

[pjb39]

What is the correct location of 11367.8.203207.ATGCACG-TCGTGCA.r1_cutadapt.fq and what the value of fastqDir?

[pjb39]

I have been doing some more testing. It turns out there is a bug after the recursive search. The code finds the filename in the right place then later looks for it in fastqDir because the person doing the downstream coding did not realize that the search was recursive. So, for now the fastq files all need to be in fastqDir. The bug will be fixed so that the recursive search is functional.

[kathrynm]

Ok, thank you for checking it out!