Hi team,
I'm wondering if there are plans to develop human removal and shotgun metagenomic preprocessing in QIIME 2.
If there are, would you be willing to share?
Thanks!
Justine
Hi team,
I'm wondering if there are plans to develop human removal and shotgun metagenomic preprocessing in QIIME 2.
If there are, would you be willing to share?
Thanks!
Justine
I am joining to the question with more generalized keyword "host". Currently I use custom script with Trimgalore and Bowtie 2, but it would be nice to have it under the hood.
One way that we have performed host read removal is by including the host references in a classification database and then filtering anything classified as such (e.g. using kraken2). I've read papers that use this mechanism and this mechanism only to remove human reads but I don't believe it's widely agreed that this is sufficient or sensitive enough, especially not if only one or two human references genomes are in the database.
Hi @jwdebelius @timanix @colinvwood ,
Yes human read filtering from shotgun metagenome data has been possible in QIIME 2 for a few releases now, as outlined by @gregcaporaso here, and we are working on further updates/improvements and a tutorial. Two approaches/three methods are already available (and combining both approaches is recommended):
The issue is not with the methods per se (not that anyone's perfect), but with the reference databases, e.g., using a suitable collection of human reference genomes instead of mapping to a single reference genome .
We will have further updates and info on this in the near future!
Thanks so much @Nicholas_Bokulich and sorry for failing at the search function! If I can help with docs/tutorials, please let me know!