Hi All,
I’m running kneaddata v0.12.0 on some libraries generated with truseq barcodes. At the trimmomatic step kneaddata tosses out roughly half of the reads of every sample. However, when I run trimmomatic independently of kneaddata I don’t observe half the reads being tossed.
Kneadata was installed by pip inside a conda environment. As is advised in other parts of the forum foward reads are run through the following sed command:
sed ‘s/ 1.*//1/g’
and reverse reads are run through the following sed command:
sed ‘s/ 2.*//2/g’
Any insight on why running trimmomatic inside of kneaddata results in so many reads being tossed would be appreciated!