cellranger module breaks with cellranger report

[pcm32]

Description of bug

Running multiqc . on a directory with cellranger reports fails due to a missing key.

File that triggers the error

PS-SCDI-022n_web_summary.html.zip

MultiQC Error log

$ multiqc .

  /// MultiQC 🔍 | v1.14

|           multiqc | Search path : /Documents/Persist-seq-unreplicated/data/pilot-fixed-rna-profiling-01/for_multiqc
|         searching | ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100% 27/27
╭────────────────────────────────────── Oops! The 'cellranger' MultiQC module broke... ───────────────────────────────────────╮
│ Please copy this log and report it at https://github.com/ewels/MultiQC/issues                                               │
│ Please attach a file that triggers the error. The last file found was: ./PS-SCDI-022n_web_summary.html                      │
│                                                                                                                             │
│ Traceback (most recent call last):                                                                                          │
│   File "/miniconda3_arm/envs/csvkit-multiqc/lib/python3.10/site-packages/multiqc/multiqc.py", line 654, in ru │
│     output = mod()                                                                                                          │
│   File "/miniconda3_arm/envs/csvkit-multiqc/lib/python3.10/site-packages/multiqc/modules/cellranger/cellrange │
│     n["count"] = self.parse_count_html()                                                                                    │
│   File "/miniconda3_arm/envs/csvkit-multiqc/lib/python3.10/site-packages/multiqc/modules/cellranger/count.py" │
│     self.parse_count_report(f)                                                                                              │
│   File "/miniconda3_arm/envs/csvkit-multiqc/lib/python3.10/site-packages/multiqc/modules/cellranger/count.py" │
│     warnings[alarm["id"]] = "FAIL"                                                                                          │
│ KeyError: 'id'                                                                                                              │
│                                                                                                                             │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
|           multiqc | No analysis results found. Cleaning up..
|           multiqc | MultiQC complete

[ewels]

Already fixed in #1841 and available in the development version.

Hoping to do a patch release v. soon, primarily to solve this bug and one with DRAGEN (just waiting for the latter, hopefully ready next week).

Thanks for reporting!

Phil

[odoublewen]

Hi @ewels, I was also wanting to use MultiQC with cellranger 7.1 output, and was also seeing the same error reported here.

Today I installed MultiQC v1.15.dev0 and tried again, and now I see a different message:

$ multiqc *html

  /// MultiQC 🔍 | v1.15.dev0 (c7b672e)

|           multiqc | Only using modules: cellranger, samtools, cutadapt, fastp
|           multiqc | Search path : /local/cell-ranger-pipeline/work/65/aa9ce2292b1180a4fb0a4b850d86d3/1_web_summary.html
|           multiqc | Search path : /local/cell-ranger-pipeline/work/65/aa9ce2292b1180a4fb0a4b850d86d3/2_web_summary.html
|           multiqc | Search path : /local/cell-ranger-pipeline/work/65/aa9ce2292b1180a4fb0a4b850d86d3/3_web_summary.html
|         searching | ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100% 3/3  
╭──────────────────── Oops! The 'cellranger' MultiQC module broke... ────────────────────╮
│ Please copy this log and report it at https://github.com/ewels/MultiQC/issues          │
│ Please attach a file that triggers the error. The last file found was:                 │
│ 2_web_summary.html                                                                     │
│                                                                                        │
│ Traceback (most recent call last):                                                     │
│   File "/local/[obfuscated]/                                                           │
│     output = mod()                                                                     │
│   File "/local/[obfuscated]/                                                           │
│     n["count"] = self.parse_count_html()                                               │
│   File "/local/[obfuscated]/                                                           │
│     self.parse_count_report(f)                                                         │
│   File "/local/[obfuscated]/                                                           │
│     "title": f"Cell Ranger count: {summary['analysis_tab']['seq_saturation_plot']['hel │
│ KeyError: 'seq_saturation_plot'                                                        │
│                                                                                        │
╰────────────────────────────────────────────────────────────────────────────────────────╯
|           multiqc | No analysis results found. Cleaning up..
|           multiqc | MultiQC complete

If it would help, I could send you one of these web_summary.html files, but I probably cannot attach it publicly.

Thank you for your tireless work maintaining and improving this important tool!

[odoublewen]

Actually, my above comment may be due to the fact that I ran cellranger on a drastically subsampled dataset -- 5000 paired reads, for a testing purposes. That might have caused cellranger to leave out the seq_saturation_plot. Let me rerun with more reads -- I will update this issue later today.

[odoublewen]

Indeed, when I ran cellranger with a full dataset, it produced web_summary.html files that were usable by multiqc 1.15.

One might argue that, as long as cellranger count runs without errors and produces a web_summary.html file, it should be valid input for multiqc. Even with a subsampled dataset, the report still has a lot of data in it that multiqc knows how to summarize. @ewels I will leave it to you to decide if you want to pursue this. Let me know if I can help.

[ewels]

Yes it should never crash like that - please create a new issue and attach a file that triggers the error (zip it if GitHub complains about the format).

Thanks!

[chris-rands]

Already fixed in #1841 and available in the development version.

Hoping to do a patch release v. soon, primarily to solve this bug and one with DRAGEN (just waiting for the latter, hopefully ready next week).

Thanks for reporting!

Phil

Hi @ewels, are you still planning the patch release or indeed multiqc 1.15 anytime soon? Thanks!

[ewels]

Sorry about this, I fell for the classic fallacy of wanting to get "one more thing" into the release, which turned out to be more difficult than expected and then I got swept into other things.

It's already at / near the top of my list, I'll try to find time to check it over and release as-is ASAP.