Hi all, I just wanted to mention that the instructions for setting up ASCAT for whole-exome sequencing are a little outdated and could be clearer in a number of places. Here are a few items that should probably be addressed: 1. In step 1, the link for downloading GC/RT files takes you to the WGS page, but I think it should take you to the WES page. Also, the links are now Zenodo links rather than Dropbox links. 2. In the code block in step 1, the

unzip ${t}_G1000_hg38.zip

command should use a file suffix of

_G1000_WES_hg38.zip

. Also, the GC correction file doesn't follow ASCAT's own specifications and includes the chr prefix for chromosomes, so the code block in step 1 should include a line to remove chr after unzipping. E.g., add

sed -i -e 's/chr//' ${t}_G1000_hg38.txt

. So, the final code block for step 1 would be:

sed -e 's/chr//' targets_with_chr.bed > targets.bed
 
for t in GC RT
do
  unzip ${t}_G1000_WES_hg38.zip

  sed -i -e 's/chr//' ${t}_G1000_hg38.txt 
  cut -f 1-3 ${t}_G1000_hg38.txt > ascat_${t}_snps_hg38.txt
  tail -n +2 ascat_${t}_snps_hg38.txt | awk '{ print $2 "\t" $3-1 "\t" $3 "\t" $1 }' > ascat_${t}_snps_hg38.bed
  bedtools intersect -a ascat_${t}_snps_hg38.bed -b targets.bed | awk '{ print $1 "_" $3 }' > ascat_${t}_snps_on_target_hg38.txt
 
  head -n 1 ${t}_G1000_hg38.txt > ${t}_G1000_on_target_hg38.txt
  grep -f ascat_${t}_snps_on_target_hg38.txt ${t}_G1000_hg38.txt >> ${t}_G1000_on_target_hg38.txt
  zip ${t}_G1000_on_target_hg38.zip ${t}_G1000_on_target_hg38.txt
 
  rm ${t}_G1000_hg38.zip
done

3. Step 2 says, "The steps below follow downloading from the Battenberg repository at the Oxford University Research Archive." But the commands actually include that part. 4. In the code block in step 2, the battenberg.zip file is not a tar archive and should simply be unzipped like

unzip battenberg.zip -d battenberg

. The final code block should be:

wget <https://ora.ox.ac.uk/objects/uuid:08e24957-7e76-438a-bd38-66c48008cf52/files/rt435gd52w>
mv rt345gd52w battenberg.zip
unzip battenberg.zip -d battenberg
cd battenberg

unzip 1000G_loci_hg38_chr.zip
cd 1000G_loci_hg38
mkdir battenberg_alleles_on_target_hg38
mv *allele* battenberg_alleles_on_target_hg38/
mkdir battenberg_loci_on_target_hg38
mv *loci* battenberg_loci_on_target_hg38/
cd ../../

5. In step 3, for consistency, the commands to copy the

targets_with_chr.bed

and

GC_G1000_on_target_hg38.txt

files should be provided. Also, the formatting could be improved throughout this section to maintain the indenting and make it clearer that there are two possible branches one should follow in the instructions at this step. 6. The "'chr'-based versus non 'chr'-based reference" part of step 3 is hard to follow and feels like it comes out of nowhere. In step 2, you download and prepare the Battenberg 1000G loci and alleles files and now suddenly we're talking about a different set of files (that seemingly do the same thing?) but haven't been downloaded yet. You're then given a command for adding the chr prefix to the (as yet undownloaded) hg19 files, even though the steps to date have been referencing hg38. There's also a comment that says "If your BAMs are ‘chr’-based, you will need to add ‘chr’", but presumably, these files should actually be prepared in a way that accounts for whichever chromosome naming convention is used throughout the Sarek pipeline. Ideally, the difference between the Battenberg vs ASCAT loci and allele files should be explained earlier, and the instructions to prepare them should be provided in a linear fashion for each (or for just the best ones if there is a consensus about that). 7. In step 4, the indentation of the Rscript should match the rest of the step.

Can I suggest that you make a PR with your suggestions please

Ok, I created this pull request with the above changes: https://github.com/nf-core/sarek/pull/2135