25/Feb/26 - Austin Hackathon notes + timeline plan

[ppreshant]

Tasks plan

• (Austin/?) Explain the pigeon module's: unitigs vs contigs vs bins plots: Edit and clarify the readme: bin/pigeon_report_guide.md included in the results in the documentation branch //
  • Make a plan for iterations on this module or updates to figures and add to an issue thread ; Check the notes from the original issue Create reads vs. contigs vs. bins comparison tool (pigeon.py) #62
• (Austin/?) Start a discussion on github/Somatem (mutual editing!) for comparing Somatem to mmlong2
• (Austin/?) Help make a plan for thorough documentation of Somatem: (readme, tutorial, examples, uses cases). Add a comment on the issue: Complete documentation in README + Wiki #63
  • checkout biobakery/maslin3 : Wiki reads like a tutorial.
• ? Make metro map showing the tools in the pipeline. Highlight things in progress ; could use this nf-metrp tool
• More things to come once we meet and discuss!

Other stuff

Note: please dump helpful links or anything else ; such as the metro map making software in this section here ; PK will try and organization them eventually!

• We are adding issues to a list called milestone to track progress here: milestone v1
• agm is trying to reproduce the taxonomic profiling from the documentation branch ; after initial file and syntax level debugging he discovered that Lemur had issues due to a recent pandas update from Feb 16th.
• To keep things working, we should peg version numbers in all the dependancies we include in the yml files of modules we build ; and whereever possible, let's use nf-core's modules which already follow this practice.
• start here!

---

Dissecting Todd's message for tasks/timeline

• 25/Feb/26, 2h hackathon: list: highest priority items; status of the code (?) + AI code review

  22/Feb/26, Todd to 2/agm, me : i’d like the both of you to meet in DH 3076 next week from 1pm - 3pm (wednesday the 25th, DH 3076) for a 2 hour hackathon of sorts, to wrap up as much as you can wrt Somatem. upon conclusion of that meeting, send me a list of the highest priority items left to finish and status of the code. ...
  use some of the meeting time to create the subway/metro map of what has been implemented as well as the parts that are still under construction. .
  ..
  please also run an AI code review tool and share the results of that in advance (cursor, claude code, gemini 3 pro/antigravity, else openai’s tool).  be sure to also prep a comparison of somatem to mmlongs2. ...

  on a separate note, I’d like Omi to be able to run Somatem or mmlongs2, as well as a 16S only pipeline (comprising of Emu and QC and outputs, nothing else). @Prashant Kalvapalle you should talk with Grace, Dongwei, and Felix about this on March 5th or 6th (after we get Somatem submitted to bioconda).

  • PK/ Understand what: status of code means?: a) things we finished in the recent past? b) ?
• 4/Mar/26, lab meeting. full operational v1: incl. detailed documentation, present design choices, 2 small examples, code review? ;

  22/Feb/26, Todd to agm, me : ... afterwards, I want a fully operational (v1) ready by Wednesday March 4th at 1pm so that we can then submit it to bioconda on Friday March 6th (in advance of the M3 meeting). ...
  also, you both will then present the Somatem code on March 4th at 1pm , so be prepared to jointly present the code, design choices etc, and go through a couple of small examples...
  and don’t sleep on the fact that thorough documentation for Somatem is going to be one of the reasons people decide to use it, so lets get a framework for that going as well (readme, tutorial, examples, uses cases, none of which should be LLM generated). once we get somatem onto bioconda, we will hopefully have a week or so before M3 (March 20th) to test and fix any issues that arise.

• 6/Mar/26, bioconda: submit to bioconda by Fri 6/Mar
• 20/Mar/26, m3/microbiome : Presenting Somatem only in full ; focus: rhea?

  (Todd to 3/Eddy, mike, me, 12/Feb/26) Somatem abstract, meaning Somatem is completely ready for primetime March 20th. lets roll out the Omi+Somatem combo later, so for now, its just the long read pipeline. featured as lightning talk or poster. which means, i’d like us to highlight the rhea integration for end to end pipeline that can reproduce the Rhea paper results (for one), (note, i’d like to feature the Omi part at the ML conf i shared yesterday, but you can tease it (omi) at the end of your talk saying that will be ready soon etc). Please confirm you can attend & present at this meeting (travel would be March 19th through the 20th or 21st).

[ppreshant]

@microbemarsh let me know if you are not able to edit the above message!

[microbemarsh]

yep looks like i can edit it!

[ppreshant]

Takaways: Update for Todd:

• We used the hackathon to get on the same page; partitioned out tasks and planned a framework for the documentation

Broad tasks for v1 (3/Mar/26) :

1. (Austin / Documentation): Starting a framework on github wiki for a tutorial style documentation with examples for 16S, metagenomic tax profiling and mag assembly using subsampled short example fastq files. inspired from biobakery/Maslin3
   • Prashant will assist with finding the files for mags ; 16S and metagenome-tax already made with Zymo-mock datasets
2. (Austin / Documentation): Polishing the documentation of the unitigs-contigs-bins comparison tool and noting how to make the plots presentable on a separate issue for Prashant
3. (Prashant / Tax-profiling outputs): Testing multi-sample krona plots and making stacked bar plots for taxonomic profiling outputs -> hand off to Dongwei's html
4. (Prashant / metro map): Start making a metro map using Austin's recommended nf-metro tool. we have a decent template already
5. (Prashant / rhea on gcloud): Run rhea through Somatem to make the plots from the paper on the google cloud while we still have it ; not urgent for v1.0-conda but before the m3 conference