How to interpret the Quality plot?

[nataradias (Natara Dias Gomes da Silva)]

Hey everyone,

I am starting my studies in Microbiome, so I don't have the experience to analyze my quality data from the forward and reverse sequences obtained after removing the primers.
As I received the files without the barcodes, I imported the paired-end-sequences to Qiime2, removed the primers using the command below (following the tutorials available), and proceeded with the quality demux of both sequences, as follows:

qiime cutadapt trim-paired
--i-demultiplexed-sequences demux-paired-end.qza
--p-cores 4
--p-front-f TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG
--p-adapter-f GGATTAGATACCCBDGTAGTCCTGTCTCTTATACACATCTCCGAGCCCACGAGAC
--p-front-r GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC
--p-adapter-r CTGCWGCCNCCCGTAGGCTGTCTCTTATACACATCTGACGCTGCCGACGA
--p-match-adapter-wildcards
--p-match-read-wildcards
--p-discard-untrimmed
--o-trimmed-sequences demux-paired-end-trimmed.qza
--verbose

After removing the primers, I had a forward sequence of 237 bases and a reverse sequence of 275 bases, with an overlap of only 38 bases. Does anyone know if this is enough to continue the analyses? Also, the Q score varied over the sequence, mainly at the end of the reverse sequence, so I don't know how to proceed with the dada2 denoise-paired. Should I use both sequences or just one of them? Still, should I trim the sequences with low Q scores or not? My concern is related to the short size of the overlapping.

Can anyone help me, please?

[colinbrislawn (Colin J Brislawn)]

Hello Natara,

Welcome to the forums!

I'm glad you have already been able to import your reads and run cutadapt.

@nataradias

Also, the Q score varied over the sequence, mainly at the end of the reverse sequence,

Unfortunately, that quality score drop in R2 is typical for an Illumina run.

@nataradias

My concern is related to the short size of the overlapping.

The overlap of 38 bases is good! But the quality drop in R2 may prevent joining.

My recommendation is to try running DADA2 paired-end with a few different trimming settings and see what preserves the most reads. If no settings work well, you can use just the forward read, as you suggested.

[nataradias (Natara Dias Gomes da Silva)]

Thanks @colinbrislawn for your advice! I'll try this!