This is a test version of Biostars. For the public version, visit https://www.biostars.org.
per sample mean coverage and standard deviation

Hello everyone

I have 18000 text files containing depth for each position in sample. each text file corresponds to 1 sample (total 18000 sample). I wanted to get the mean coverage, standard deviation and total count of positions per sample in a single output file. I was just wondering if there is an easy way to do it? depths were calculated using samtools depth input.bam. all the text files looks like this...

sample_name   chromosome   position     depth

so the desired output is..

sample1  mean_depth   standard_deviation   total_number_of_positions
sample2  mean_depth   standard_deviation   total_number_of_positions
sample1  mean_depth   standard_deviation   total_number_of_positions
bash awk
• 2,182 views • link
updated by husensofteng husensofteng • written by ijlal.hyder2012 • ijlal.hyder2012

this is a different question so I thought of asking it in a different post. there I wanted average depth per position across all the samples. here I need per sample average depth, sd and counts for total number of positions.

• 0 views • link
written by ijlal.hyder2012 • ijlal.hyder2012

You can probably calculate that using some variation of datamash solution that was posted by @cpad0112 in last question. Tinkering with things is a great way to learn.

You should also validate answers for your past questions, if they helped you solve the issue (green check mark besides answers). You can accept more than one if they all work.

• 0 views • link
written by GenoMax GenoMax

1 answer

If I understand correctly, you want the overall quants across all positions per sample. In such case, assuming that the depth values are in the fourth column and all .txt files are in the directory:

for sample_file in *.txt; do 
   awk 'BEGIN{OFS="\t"}{x+=$4; y+=$4^2}END{print $1,x/NR,sqrt(y/NR-(x/NR)^2),NR}' $sample_file >> output.txt; 
done
• 0 views • link
written by husensofteng husensofteng

Log in to answer this question.

More posts like this