Using cutadapt with a paired collection produces an accurate cutadapt report, with a full list of adapter sequences trimmed; however, the sequence output is an empty list. I selected “Multiple output: create a separate file for each adapter trimmed (default: all trimmed reads are in a single file)” but still only receive an empty list as the output.
Please note that when using cutadapt with a single read file, the report is generated correctly, and the multiple outputs are generated correctly as a list. Is there an optimal way to specify cutadapt paired read inputs to generate the correct output?
Unfortunately TrimGalore and Trimmomatic cannot perform the cutadapt function of separating the reads by internal adapter, while retaining the “untrimmed” sequence. cutadapt can accomplish this function easily in a unix/linux environment, but fails on Galaxy.
I’ve been having a similar problem. When I use a fastq file as the input, I got prefect output. But when I convert the exact same file to fasta format and run the tool again with the same parameter, I got a great report but an empty output list. Some of my data has to be converted to fasta first for upstream analysis so it’'s been troubling me. I switched to command line cutadapt tool which worked for me, but really hope someone can explain this report without output issue.
Have a list of fastq pairs in a nested collection.
Run Flatten Collection to simplify the data structure. This will rename the files – and you want this to happen.
Now, run Cutadapt on the flattened data.
Please give that a try. If this still fails or produces unexpected results, would you please share back an example of the problem with the smallest input data possible?
I’d like to troubleshooting what is going on and that may involve some developers, so the data needs to be small and public, and isolated in a dedicated history.