normalize fastq files prior to merging and mapping

[mpoelchau]

• try installing bbnorm from the bbmap package. https://sourceforge.net/projects/bbmap/
• guide on how to use bbnorm: https://jgi.doe.gov/data-and-tools/bbtools/bb-tools-user-guide/bbnorm-guide/
• example command for paired-end files: bbnorm.sh in1=read1.fq in2=read2.fq out1=normalized1.fq out2=normalized2.fq target=100

if this works, then it will probably replace the post-mapping downsampling component.

open questions:

• what target parameter to use?
• should we normalize multiple times - first, on individual fastq files, then merge them, then normalize again?
• how much time and compute does this take with an average set of fastq files?

[HsiuKangHuang]

Do "in1=read1.fq" and "in2=read2.fq" mean the output fastq files from the step trimmomatic? Is this normalizing step before the aligning step?

[mpoelchau]

Yes!