Should I cluster my reference sequences to 97% for my classifier?

Library Plugin Support
, ,
1

Hello Devon and Mike,
One question. If I am trying to build a classifier for 97% OTU, should I use --p-perc-identity 0.97 command during dereplication step as explained in here?
thanks!
Lautaro

2

I’d recommend not clustering your reference data unless you are trying to limit memory and storage use.

You’re likely to get worse classifications too, see our RESCRIPt preprint for some insight.

2 Likes