Sequence depth normalization after woltka classification

[UniAlberta]

Hi there, Im using woltka classify on metagenomic short reads. I got the output as RPK which gave me the copy number of each gene in each sample. Although, Id like to normalize the sequencing depth in each sample as well. Is there any command to use for normalizing sequence depth on the RPK output table?

Thanks

[qiyunzhu]

Hello @UniAlberta Yes you can do that. See the instruction. Basically, you need:

woltka normalize -i rpk.biom --scale 1M -o tpm.biom

[UniAlberta]

Hi and thanks for your reply. So you are suggesting to convert RPK to TPM?
I use this workflow to create RPK outputs in tsv format.
#####woltka classify
--input $SAM_INPUT_DIR/
--coords $DB_DIR/proteins/coords.txt
--map $DB_DIR/function/uniref/uniref.map.xz
--names $DB_DIR/function/uniref/uniref.name.xz
--names $DB_DIR/function/kegg/ko.name
--map $DB_DIR/function/kegg/ko.map.xz
--rank uniref,ko
--sizes .
--scale 1k
--digits 3
--to-tsv
--output $OUTPUT_DIR

How can I add "woltka normalize -i rpk.biom --scale 1M -o tpm.biom" to this workflow so that the final output is normalized by both gene length and sequence depth? I appreciate your help.

[qiyunzhu]

@UniAlberta After running the command you posted, you will get two output files: uniref.tsv and ko.tsv. Then you can run the command I suggested on each of them, like woltka normalize -i uniref.tsv --scale 1M -o uniref.norm.tsv

[UniAlberta]

Thank you very much!