Replicating QIIME2 parameters in ampliseq

[carsharp]

I am losing a lot of reads when I do quality-filtering of my 16S MiSeq run data that I got back from a sequencing facility. See the settings I am using below:

nextflow run nf-core/ampliseq
-r 2.3.2
-profile singularity
--input './samplesheet_16S_Carly.tsv'
--FW_primer GTGYCAGCMGCCGCGGTAA
--RV_primer GGACTACNVGGGTWTCTAAT
--metadata "./Sharp_16S_Metadata.txt"
--outdir "./nextflow_dir_2"
--ignore_empty_input_files
--retain_untrimmed \

When I contacted the sequencing facility, they sent the following recommendation for the DADA2 plugin in QIIME2 in which they were able to retain a healthy fraction of non-chimeric reads:

Dada2 denoise-single method requires 2 parameters that are used in
quality filtering:

1. --p-trim-left: trims off the first m bases of each sequence
2. --p-trunc-len: which truncates each sequence at position n

qiime=qiime2-2022.11

source ~/.bashrc
conda activate $qiime

qiime dada2 denoise-paired
--i-demultiplexed-seqs paired-end-demux.qza
--p-trim-left-f 12
--p-trim-left-r 12
--p-trunc-len-f 200
--p-trunc-len-r 200
--o-table table.qza
--o-representative-sequences rep-seqs.qza
--o-denoising-stats denoising-stats.qza
--p-n-threads 4
--p-max-ee-f 2
--p-max-ee-r 2

I would like to do my own analysis in ampliseq; is there a way that I can replicate these trim-left-f 12/trim-left-r 12 and trunc-len-f 200/trunc-len-r 200 parameters in ampliseq?

[d4straub]

Hi there,
I would be very careful of the results when there is no justification why --p-trim-left-f/--p-trim-left-r should be used for that data. The other settings seem ok.
Anyway, here is the breakdown:

--p-trunc-len-f 200
--p-trunc-len-r 200

use --trunclenf 200 --trunclenr 200

--p-max-ee-f 2
--p-max-ee-r 2

thats already default

--p-trim-left-f 12
--p-trim-left-r 12

is more tricky, here we have to modify via a config file the process parameters, because the pipeline doesnt offer to modify them directly. So create a file, lets call it trim.config, that contains:

process {
    withName: DADA2_FILTNTRIM {
        ext.args = [
            'maxN = 0, truncQ = 2, trimRight = 0, minQ = 0, rm.lowcomplex = 0, orient.fwd = NULL, matchIDs = FALSE, id.sep = "\\\\s", id.field = NULL, n = 1e+05, OMP = TRUE, qualityType = "Auto", maxEE = c(2, 2), trimLeft = 12, minLen = 50, maxLen = Inf, rm.phix = TRUE'
    }
}

and the run your pipeline with the addition -c trim.config to apply those settings. The only change from default is here trimLeft = 12 [potentially it needs to be trimLeft = c(12,12)]. I didnt test it but it might work, if not there will be an error.
Documentation about params that can be used for that process can be found in https://rdrr.io/bioc/dada2/man/filterAndTrim.html.

[carsharp]

Thanks so much for your response! I was able to implement the QIIME2 parameters in ampliseq per your suggestions. I appreciate your note of caution regarding using the--p-trim-left-f/--p-trim-left-r parameters. For some more context, the sequencing facility suggested this because they observed a quality drop/noise at the beginning of the reverse sequences, which they said they occasionally observe in their 16S miseq runs due to tendencies for MiSeq 2×250 bp amplicon runs to show lower base quality in the first cycles of Read 2, especially with low-diversity 16S libraries. According to them this isn’t uncommon and is handled by their standard primer/leading-base trimming and denoising.

[d4straub]

For 16S reads that were produced with MiSeq 2x250bp I havent experienced problems in the first cycles (after the primer) yet. Potentially they are adding too less PhiX to diversify the library. Anyway, glad it works for you now!
I close that issue but feel free to open it or another time when the need arises.