UMI-Tools: support extract command

[mschubert]

Description of bug

The regexes provided in the UMI-Tools module

MultiQC/multiqc/modules/umitools/umitools.py

Lines 105 to 114 in d26af5e

	regexes = {
	"total_umis": r"INFO total_umis (\d+)",
	"unique_umis": r"INFO #umis (\d+)",
	"input_reads": r"INFO Reads: Input Reads: (\d+)",
	"output_reads": r"INFO Number of reads out: (\d+)",
	"positions_deduplicated": r"INFO Total number of positions deduplicated: (\d+)",
	"mean_umi_per_pos": r"INFO Mean number of unique UMIs per position: ([\d\.]+)",
	"max_umi_per_pos": r"INFO Max. number of unique UMIs per position: (\d+)",
	"version": r"# UMI-tools version: ([\d\.]+)",
	}

do not match the UMI-Tools log file output when using --extract_method regex (exact command via nf-core/rnaseq here). The corresponding log file looks the following (full file attached):

2024-01-25 00:20:39,419 INFO Parsed 10000000 reads
2024-01-25 00:20:40,180 INFO Input Reads: 10077011
2024-01-25 00:20:40,180 INFO regex does not match read1: 9215286
2024-01-25 00:20:40,180 INFO regex matches read1: 861725
2024-01-25 00:20:40,180 INFO regex does not match read2: 752463
2024-01-25 00:20:40,180 INFO regex matches read2: 109262
2024-01-25 00:20:40,180 INFO Reads output: 109262

MultiQC version: 1.14 (but regexes are the same in latest)
UMI-Tools version: 1.1.4 (latest)

File that triggers the error

RMGI_wt_250_48h-3.umi_extract.log

MultiQC Error log

No error, but UMI-Tools panel is absent.

Before submitting

• I have read the troubleshooting documentation.
• I am using the latest release of MultiQC.
• I have included a full MultiQC log, not truncated.
• I have attached an input file (.zip if necessary) that triggers the error.

[vladsavelyev]

Thanks for the issue! Indeed, The UMI-Tools MultiQC module only supports the dedup command.

I added some basic support for extract now with #2296, let me know if you have ideas of any better visualisations for it apart from a simple table.

[mschubert]

This is great, thanks!

For the vis, I'd probably suggest a stacked bar graph with colors: (1) mismatched read1, (2) matched read1 but mismatched read 2, (3) matched both reads; total of these is the input read number

[vladsavelyev]

Should be done with #2296