benn888/PRONAME ? reverse-engineered prompt

Build me a simple bioinformatics pipeline for Nanopore metabarcoding data that I can run from Docker.

I want it to take a folder of basecalled FASTQ files, then run through four clear steps, first import and optionally trim adapters and primers, then filter reads by length and quality and separate duplex from simplex reads, then cluster and polish the reads to make high accuracy consensus sequences and remove chimeras, and finally do taxonomic classification on the results. It should be easy to run step by step from the command line, with helpful help text for each script and sensible defaults.

Please make it work on both amd64 and arm64, include the needed scripts and reference database in the container, and let me mount a local data folder so inputs and outputs are easy to access. If possible, also support using the GPU during the polishing step. I’d like the workflow to feel straightforward for someone with basic command line experience, and to produce outputs that can later be used in QIIME2 or a phyloseq object if needed.