Hi,
Thanks for the work on this tool.
I really appreciate the addition (or maybe it was always there and I hadn't spotted it?) of the split-by-low-quality function. I'm implementing this in some current analyses I'm running, but I'm running into a downstream issue with methylation tag data no long matching the trimmed (and/or split) read coordinates.
(EDIT: a related question/clarification: is the order of operations always: 1. split-by-low-quality; then 2. filter based on min length and mean q score?)
For context: This is nanopore data, and I've generated methylation data via dorado with a few of the available models, converted bam to fastq keeping the tags in the sequence headers (samtools fastq -t '*' ...), and then am processing via chopper. However, if I include any trimming (or splitting of reads via split-by-low-quality), the methylation tag (MM and ML) data no longer correspond correctly to the chopped sequence coordinates.
Would it be possible to implement the option to parse MM and ML tag data, and then regenerate these data correcting the coordinates based on any trimming or splitting that chopper has done?
I imagine parallel methylation analyses will be increasingly popular, and I also see a lot of value in both quality trimming ends of reads and also in the split-by-low-quality approach. It would be great to have just the one quality filtered and trimmed (and/or split) data set for downstream use (i.e. rather than having one non-trimmed/split set for methylation analyses and a separate trimmed/split set for all other analyses).
Cheers!
Mike.
Hi,
Thanks for the work on this tool.
I really appreciate the addition (or maybe it was always there and I hadn't spotted it?) of the
split-by-low-qualityfunction. I'm implementing this in some current analyses I'm running, but I'm running into a downstream issue with methylation tag data no long matching the trimmed (and/or split) read coordinates.(EDIT: a related question/clarification: is the order of operations always: 1. split-by-low-quality; then 2. filter based on min length and mean q score?)
For context: This is nanopore data, and I've generated methylation data via dorado with a few of the available models, converted bam to fastq keeping the tags in the sequence headers (
samtools fastq -t '*' ...), and then am processing via chopper. However, if I include any trimming (or splitting of reads viasplit-by-low-quality), the methylation tag (MM and ML) data no longer correspond correctly to the chopped sequence coordinates.Would it be possible to implement the option to parse MM and ML tag data, and then regenerate these data correcting the coordinates based on any trimming or splitting that chopper has done?
I imagine parallel methylation analyses will be increasingly popular, and I also see a lot of value in both quality trimming ends of reads and also in the split-by-low-quality approach. It would be great to have just the one quality filtered and trimmed (and/or split) data set for downstream use (i.e. rather than having one non-trimmed/split set for methylation analyses and a separate trimmed/split set for all other analyses).
Cheers!
Mike.