Incorrect sequencing yield calculated by bcl convert module when using asymmetric read lengths

[BigTA78]

Description of bug

When invoking the bclConvert module on a sequencing run where read 1 and read 2 read-lengths differ (e.g. a 10X Genomics scRNA run, where R1=28 cycles and R=90), yields are incorrectly calculated.

The assumption is 2 * R1 * # clusters which underestimates yield. This leads to over-estimate of % bases >Q30, with values over 100% (see attached report bundled with bclConvert output folder). Actual calculation should be # cluters * (R1+R2)

File that triggers the error

bclConvert_example.zip

MultiQC Error log

$ multiqc -m bclconvert home/user/bclConvert_example/ -n /home/user/bclConvert_example/multiQC.html

  /// MultiQC 🔍 | v1.14.dev0

|           multiqc | Only using modules: bclconvert
|           multiqc | Search path : /home/user/bclConvert_example
|         searching | ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100% 13/13
|        bclconvert | 1 lanes and 11 samples found
|           multiqc | Compressing plot data
|           multiqc | Report      : /home/user/bclConvert_example/multiQC.html
|           multiqc | Data        : /home/user/bclConvert_example/multiQC_data
|           multiqc | MultiQC complete

[ewels]

x-ref #1697 and #1701

Maybe @andrei-seleznev might have some insight here? 🙏🏻

[andrei-seleznev]

Should be straightforward enough to grab R1 and R2 length from RunInfo.xml and calculate appropriately - I can make this change next week and PR!

[ewels]

Thanks @andrei-seleznev !