search

USDA-ARS-GBRU / itsxpress

Software to trim the ITS region of FASTQ sequences for amplicon sequencing analysis
Other
9 stars 9 forks source link

A question #14

Closed chaibenl closed 4 years ago

chaibenl commented 4 years ago

Does ITSxpress trim paired-end reads that can't be merged because of lack of overlap?

arivers commented 4 years ago

No, it does not trim or return reads that it cannot merge.

chaibenl commented 4 years ago

What if I concatenate R1 and R2 and treat it as if it contains single end reads?

arivers commented 4 years ago

Could you provide more background about your sequences, sequencing type and what you plan do with the reads after trimming?

I'd like to provide advice that makes sense for your scientific question.

chaibenl commented 4 years ago

Thank for the help! I have 2X150 Illumina Miseq sets of Fungal ITS1. I found most of the reads can't be merged due to lack of overlap. I plan to run through dada2 or deblur and then classify the reads after rRNA trimming. The goal is to generate a feature table of Fungal taxonomic composition for the sampling environments. Hope that helps! Thank you.

arivers commented 4 years ago

The current version of ITSxpress uses pretty stringent BBmerge merging. You may want to do a test of the merging rate using Pear with defaults, BBmerge with defaults and BBmerge using the parameter maxmismatches=100. if your merging rate improves with those can change a setting on ITSxpress to get more merging.

Ultimately the reads need to merge to use Dada2 because Dada2 uses the overlap to estimate the error rate. Dada2 has a single-ended option but feeding merged reads into it violates the assumptions of the error model so you shouldn't do that.

chaibenl commented 4 years ago

Thank you for the explanation! I appreciate it!

Sent from my iPhone

On Sep 27, 2019, at 2:59 PM, Adam Rivers notifications@github.com wrote:

The current version of ITSxpress uses pretty stringent BBmerge merging. You may want to do a test of the merging rate using Pear with defaults, BBmerge with defaults and BBmerge using the parameter maxmismatches=100. if your merging rate improves with those can change a setting on ITSxpress to get more merging.

Ultimately the reads need to merge to use Dada2 because Dada2 uses the overlap to estimate the error rate. Dada2 has a single-ended option but feeding merged reads into it violates the assumptions of the error model so you shouldn't do that.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.