https://nf-co.re logo
Join Slack
Powered by
Trying to create a reference for ASCAT using the 5...
# sarek
j
Trying to create a reference for ASCAT using the 5 steps detailed here. I clicked the link to get the 4 files from this ASCAT page and noted: 1.) There is now an option for a whole exome references and 2). the file names have also been changed for the linked WGS and new WES so the code you provided does not run out of the box. I got things running up until step 4 with the line 
rm 1kg.phase3.v5a_GRCh38nounref_allele_index_chr23.txt
which does not seem to exist in the reference files I downloaded. Do I need something else?
t
You want an ASCAT reference for WES, right? I found the instructions unnecessarily complicated -- they would require rebuilding the reference every time you are using a new WES kit. I think it makes more sense to run the standard files with the targets BED file, so I did that. The major things to keep in mind when setting up reference files like that is to get the directory structure just right (sarek is looking for certain names in the directories and subdirectories, and some of the files (but not all) need to have the contig names prefixed with 
chr
to be compatible with iGenomes.
j
Thanks for the reply Thomas. Could you give me a hint as to the directory structure required? I keep getting an error from ASCAT and it seems to be caused by an unexpected directory structure but I'm not sure what I'm doing wrong.
I answered my own question after submitting 20 jobs to my HPC: For anyone using ASCAT with WHOLE EXOME SEQUENCING, here is what I did: 1. I downloaded all four sets of files from the links given on this page. I am using hg38. 2. Rename "[GC/RT]_G1000_WES_hg38.zip" to "[GC/RT]_G1000_hg38.zip". ASCAT will use the folder name as the prefix for the file names inside. However, the files they provide inside the zip folders do not have "WES" in the names so you must remove it from the zip folder name. 3. The "G1000_[alleles/loci]_WES_hg38.zip" files contain another folder "G1000_[alleles/loci]All_hg38". This will break ASCAT. Unzip the folder and take all of the files out of the "G1000_[alleles/loci]All_hg38" folder and rezip them into "G1000_[alleles/loci]_hg38.zip". Again, NO "WES" in the zip folder name and the files must be top level inside the zip, no directory structure. 4. While you have the "loci" files unzipped, check if you need to add the "chr" prefix based on your BAM files. If so, follow the instructions down at the bottom of the reference download page (here). 5. ASCAT needs information on sample sex. If you are starting from step "variant_calling" you must have the "sex" column in your samples.csv file. I did not trim the reference files to my BED regions but rather ran with the 
--intervals
option in Sarek. No errors were thrown but I'm not sure if this affects results as this is my first time doing CNV analysis this way. Whew!
🙌 2
t
Yes, that's the correct way, glad you figured it out, and thank you for the clear documentation. There are is more than one way to this, since the pipeline can either take a ZIP file (with the expected structure inside) or the directory itself (i.e. rezipping is optional). Instead of renaming, I have used symlinks to make the process transparent.
c
I also recently encountered a number of issues/points of confusion when trying to follow the current documentation and ultimately created a pull request to clarify the process. Hopefully it gets approved.
👍 1
Open in Slack
PreviousNext
Powered by