Feature classifier

[zhang_sonic]

Hello
I am trying to train my classifier for taxonomy.
I used this command below to denoising my data

qiime dada2 denoise-paired
–i-demultiplexed-seqs collabration-demux-paired-end.qza
–p-trim-left-f 6
–p-trim-left-r 5
–p-trunc-len-f 300
–p-trunc-len-r 223
–p-n-threads 4
–o-table collabration-table.qza
–o-representative-sequences collabration-rep-seqs.qza
–o-denoising-stats collabration-denoising-stats.qza &

and this is my primers information
Target
341F - 806R
Forward Primer(341F)
CCTAYGGGRBGCASCAG
ReversePrimer(806R)
GGACTACNNGGGTATCTAAT
ReadLength
300bpPE

My question is when I am trying to run the feature classifier part using the command belove, what should I put for the p-trunc-len part? The way I do, is that correct?

qiime feature-classifier extract-reads
–i-sequences 99_otus.qza
–p-f-primer CCTAYGGGRBGCASCAG
–p-r-primer GGACTACNNGGGTATCTAAT
–p-trunc-len-f 300 *
** --p-trunc-len-r 223 *
–o-reads collaborator-ref-seqs.qza

[Mehrbod_Estaki]

Hi @zhang_sonic,

@zhang_sonic

My question is when I am trying to run the feature classifier part using the command belove, what should I put for the p-trunc-len part? The way I do, is that correct?

Have you had a chance to look through the tutorial covering Training feature classifiers? In particular the section regarding extracting reads should be of help. Here is what the first Note box says:

"For classification of paired-end reads and untrimmed single-end reads, we recommend training a classifier on sequences that have been extracted at the appropriate primer sites, but are not trimmed."

So in your case you would simply include your primers but don't trim them.
Hope that helps!

[Mehrbod_Estaki]

Hi @zhang_sonic,
Just wanted to point you towards this pre-trained classifier that has been extracted from the same region as you mention that you could use if you’d rather not train your own.

[zhang_sonic]

i see. Thanks for your help. Let me have a look