Arjun Krishnan
ML/AI methods & tools for using massive public data collections to gain insights into complex disease mechanisms.
Associate Professor & Group leader thekrishnanlab.org at the Dept. of Biomedical Informatics at CU Anschutz.
- I'm looking forward to re-teaching: Rethinking Data Analysis — A researcher’s guide to avoiding missteps and misuse This is an advanced short course on developing a mental toolkit for rigorous practice & critical consumption of statistical data analyses. 🧵 1/4
-
View full thread3/4 As a result, most students piece together a mental model of acceptable, standard, or "best" practices in their field from shards of information gathered from mentors, peers, and published papers.
- 4/4 This advanced short course formalizes the instruction of these ideas. The goal is to: 1) Discuss common misunderstandings & typical errors in the practice of statistical data analysis. 2) Provide a mental toolkit for critically thinking about statistical methods & results. Feedback welcome 🙌🏼
- 2/4 Statistical inquiry, data analysis, and visualization are immensely powerful, but many of the ideas underlying them are nuanced and unintuitive. Unfortunately, these ideas—and the skills needed to apply them to real problems and datasets—are rarely taught in statistics or data-analysis courses.
- Our Perspective article on Computational Strategies for Cross-Species Knowledge Transfer is now published in @natmethods.nature.com! This was a collab b/w @krishnanlab.bsky.social & @fishevodevogeno.bsky.social, led by the amazing Hao Yuan @yhbioinfo.bsky.social. 🧵 www.nature.com/articles/s41...
-
View full thread9/10 By embracing data-driven, evolution-agnostic approaches, we believe that the field can accelerate discoveries in both common and rare diseases, improving model organism selection and ultimately paving the way for more reliable therapeutic interventions.
- 10/10 Big thanks to NIH/NIGMS, NSF, & @simonsfoundation.org for funding this work! We welcome feedback from the community! 🙌 #Bioinformatics #TranslationalResearch #OpenScience
- 7/10 Kudos to resources like @geneontology.bsky.social , @monarchinitiative.bsky.social, @alliancegenome.bsky.social, & @bgee.org for grounding so much data & knowledge in this space in structured formats. These & many others are included in our catalog ☝🏽
- 8/10 Key future directions we highlight: - Capturing specific facets of complex diseases - Building networks for more species & contexts - Automated ontology/knowledge graph construction - Better benchmarking for cross-species single-cell methods - Leveraging non-traditional research organisms
- 6/10 We provide detailed resources to help computational & wet-lab researchers find, improve-upon, and apply appropriate methods: 📊 Supp Table 1: Comprehensive catalog of methods (name, category, input/output, data types) 📚 Supp Table 2 & Note: Valuable datasets for cross-species work
- 3/10 Our article covers methods that tackle 4 key questions in cross-species research: 1. Predicting function/disease-gene relationships across species 2. Identifying agnologous molecular components 3. Inferring perturbed transcriptomes across species 4. Mapping agnologous cell types and states
- 4/10 Traditional approaches rely heavily on homology. But shared ancestry ≠ shared function & vice-versa. Here, we introduce the concept of Agnology, which embraces this complexity: "agno-" = unknown/not known, reflecting data-driven functional equivalence regardless of evolutionary origin.
- 2/10 #ResearchOrganisms like 🐭 & 🐟 are crucial for studying genes, functions, cell types, & disease. But translating findings to 👨⚕️ is tricky. We explore data-driven methods to bridge the gap & introduce the concept of "agnologs" — functional equivalents identified independent of evolutionary origin.
- [Not loaded yet]
- Same from @cp-cellreports.bsky.social: Dec 24th, 25th, & 27th. Closed/removed on Dec 30th!
- Congratulations! Kudos to @richabdill.com & @samanthagraham.bsky.social for leading this huge project! Thanks for bring us onboard! Mansooreh Ahmadian & Parker Hicks lead the part of the work on inferring study annotations from unstructured metadata and text from the linked publications.
- Today we report a new compendium of human gut microbiomes with >168,000 samples By analyzing this massive dataset, we discovered distinct microbiome patterns across the globe, and show we can predict where a person lives just from their gut bacteria Now out in Cell: www.cell.com/cell/fulltex...
- I have created a starter pack for AI for Science. Let me know if you would like to be added. go.bsky.app/FPgU6Ptat://did:plc:t6dn56uvxz55kmc4y3uemcme/app.bsky.graph.starterpack/3lf34vwlu532z
- 🖐🏼
[Not loaded yet]- We just use the #papers-articles channel our group’s Slack workspace.
- A favorite! Interestingly, Goodhart stated (in 1975): “Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.” Marilyn Strathern generalized it in 1997 to its famous version👇🏽 pmc.ncbi.nlm.nih.gov/articles/PMC...
- “When a measure becomes a target, it ceases to be a good measure." - Goodhart’s Law en.m.wikipedia.org/wiki/Goodhar...
- Monday motivation from the wonderful @pracheeac.bsky.social: “[I hate the term] incentive structure… if you ask people to do a thing that they perceive to be against those processes and principles, well, you can't ask that — that it's somehow unacceptable…
- … And what I hear is that we need smthing to be good for ourselves, our careers, & our personal advancement in order to do smthing that's good for science[/world]. And I just don't believe that's true.” Love it: what’s the best thing to do, the right thing to do? podcasts.apple.com/us/podcast/o...
- Incidentally, on ping with this discussion… bsky.app/profile/harm...
- [Not loaded yet]
- 🖐🏽
- As a fan of the super @nightsciencepod.bsky.social (highly recommend it!), I enjoyed listening to the latest episode of another favorite — Work Life — where @adamgrant.bsky.social talks to Nathan Myhrvold about invention and creativity! podcasts.apple.com/us/podcast/w...
- [Not loaded yet]
- 🖐🏽
- [Not loaded yet]
- 🖐🏽
- [Not loaded yet]
- 🖐🏽
- [Not loaded yet]
- [Not loaded yet]
- This is important. Please count need in.
- [Not loaded yet]
- Yes. I use this format.
- The world doesn’t need one more blog, and yet I started one and plan to write semi-regularly! compbiologist.substack.com/p/computing-...
- [Not loaded yet]
- The topics and resources look great, Michael. I teach data & statistics communication as a small part of my "Gaps, missteps, and errors in data analysis" course. It'll be great to learn more about how you go about it. I'll take a closer look at the website soon. Thanks for sharing!
- Very much looking forward to #MLCB2023, meeting folks & learning a lot! Do checkout our PMLR poster, led by Renming Liu: Open Biomedical Network Benchmark: A Python Toolkit for Benchmarking Datasets with Biomedical Networks Preprint: doi.org/10.1101/2023... Code: github.com/krishnanlab/...
- Machine Learning in Comp Bio #MLCB2023 is Thursday/Friday this week, great line up of contributed and invited talks (sites.google.com/cs.washingto...) which we'll be livestreaming (link to follow!), along with poster flash videos.
- Everything you said + Google Scholar citation & author alerts + Semantic Scholar email digests.
- I strongly recommend getting a paid account on Interfolio, asking your letter writers to upload their letter there ONCE, and you sharing the letters directly from there for all your applications, even those that request referee emails!
- When applying for tenure track jobs how many did you apply for? I feel bad asking for so many letters of recommendation, but then I hear about people applying to 60+ places... #AcademicSky
- [1st post!] Checkout this comprehensive review & perspective on the "Current and future directions in network biology" arxiv.org/abs/2309.08478. A bunch of us got to lead the section on "Machine learning on networks". Kudos to the whole community and to Dr. Tijana Milenković for the leadership!
