bokulich-lab/RESCRIPt

How to train the classifier with mixed forward primers?

Closed this issue · 2 comments

Hi Colin,

I hope you are doing well! Thanks for sharing this useful tool!

I am analysing V1-V2 16S rRNA sequence data. I want to use qiime feature-classifier extract-reads to extract reads and train a classifier.
However, this data has mixed primers. I came across this one: How to train the classifier with multiple reverse primers? . Could you please help me to have a look?

V1-V2 MiSeq primers
Forward: These primers are mixed at a 4:1:1:1 ratio (28F-YM is the 4)
28F-YM:        GAGTTTGATYMTGGCTCAG 
28F-Borrellia: GAGTTTGATCCTGGCTTAG 
28FChloroflex: GAATTTGATCTTGGTTCAG 
28F-Bifdo:     GGGTTCGATTCTGGCTCAG


28F-YM vs 28F-Borrellia           ==   3
28F-YM vs 28FChloroflex           ==   4
28F-YM vs 28F-Bifdo               ==   4
28F-Borrellia vs 28FChloroflex    ==   4
28F-Borrellia vs 28F-Bifdo        ==   4
28FChloroflex vs 28F-Bifdo        ==   6

(19-4) differences / 19 bp length == 78.95% similar

Another problem is these primers are mixed at a 4:1:1:1 ratio (28F-YM is the 4)? Should I use 28F-YM & --p-identity 0.7/0.8?

Any advices would be highly appreciated! Thanks!

Kind regards,
Birong

hello, please post all user support questions to the QIIME 2 forum. Thank you.

Sorry for this, I post it to QIIME 2 forum yesterday: How to train the classifier with multiple mixed forward primers? . Could you please have a look? Many thanks!