Reposted by Amy D WillisCarl T. Bergstrom @carlbergstrom.com · 21/07/202613. Hence our dismal conclusion: rather than finding that LLMs free us to do a better job of what we were doing before they came along, they shift scientific incentives (and the playing field of academic competition) in ways that compel us to do more and more, faster and faster, less and less well. 251005309
Amy D Willis @amydwillis.bsky.social · 22/07/2026Understood -- thanks for sharing, @erikvannimwegen.bsky.social !! That's a helpful perspective. 000
Amy D Willis @amydwillis.bsky.social · 22/07/2026Thanks so much for sharing, @dutchscientist.bsky.social !! Much appreciated. 110
Amy D Willis @amydwillis.bsky.social · 22/07/2026Super helpful to know -- thanks!! I appreciate you taking the time, Alex!! @acritschristoph.bsky.social 000
Amy D Willis @amydwillis.bsky.social · 20/07/2026Many thanks to all for sharing, and feel free to email/DM me if you don't want to answer publicly. These chats really help me understand what you all need in terms of methods. So, thank you 🙏🏻🙏🏻 ❤️, Amy 7/ 100
Amy D Willis @amydwillis.bsky.social · 20/07/2026Without exaggeration, uncertainty quantification is my livelihood -- I'm a statistician, that's what we do. So, I know what these numbers mean to me... but what do they mean to you all, dear bacteria/microbiome friends, when you produce and read numbers that imply such confidence? 6/ 130
Amy D Willis @amydwillis.bsky.social · 20/07/2026(And yes, I am familiar with the debate about this paper and the Author Correction.) 5/ 100
Amy D Willis @amydwillis.bsky.social · 20/07/2026and almost all of these splits have >99% posterior probability (Fig 2 of the same paper) 4/ 100
Amy D Willis @amydwillis.bsky.social · 20/07/2026Here's what I'm talking about: Almost all of these splits have >90% bootstrap support (Fig 1 of www.nature.com/articles/s41...) 3/ 100
Amy D Willis @amydwillis.bsky.social · 20/07/2026Almost all of these numbers are >90%, >95%, >99%... That level of confidence may makes sense for more recent evolution, but the same confident numbers appear when we are talking divergences from billions and billions of years ago. 2/ 100
Amy D Willis @amydwillis.bsky.social · 20/07/2026Genuine question for folx who look at bacterial phylogenies: How do you interpret the bootstrap support and posterior probabilities that you get on your tree estimates? Do you take them seriously? Do you think: "Wow, all 99%, that must be a reliable tree!"? 1/ 4109
Amy D Willis @amydwillis.bsky.social · 25/06/2026I'm hiring a #postdoc 🥳 and applications are open 🤩 Candidates with interests in EITHER statistical methodology or microbial ecology / microbiome are welcome to apply. The position / projects will be tailored to the candidate 🌟📈🎤 Thank you for sharing widely! apply.interfolio.com/188571apply.interfolio.com Apply - Interfolio {{$ctrl.$state.data.pageTitle}} - Apply - Interfolio 02021
Amy D Willis @amydwillis.bsky.social · 16/06/2026I'm hiring a #postdoc!! 🥳 Official ad is making its way through HR but If you have interests in statistical methodology OR microbial bioinformatics, please reach out via my UW email. I'll share the description with you when it's available. 🔥😻 02521
Amy D Willis @amydwillis.bsky.social · 18/05/2026This isn't a full-picture benchmark against all other methods BUT I am rather pleased and thought I would share 🥳 A great argument for radEmu's robust-centering approach to identifiability 🔥 and valid error rate control 😻 Those score tests ran overnight, but were definitely worth waiting for 🎤 030
Amy D Willis @amydwillis.bsky.social · 18/05/2026radEmu took the cake today. I'm looking at a dataset where the truth is known (a complex spike-in, spiked-in at varying intensities). All taxa should be unchanging in abundance except for the spike-ins... True "positives": 6 / 1505 radEmu score tests identifies 7 CLR + linear model identifies 517 170
Amy D Willis @amydwillis.bsky.social · 20/03/2026Are you doing prediction using microbial community data as input? Please report accuracy as *average percentage error* of the prediction relative to the measurement. I'm not interested in their correlation -- I want to know how many orders magnitude you're off by. Thx! #microbiome #deeplearning 040
Amy D Willis @amydwillis.bsky.social · 16/12/2025Thx for checking! Co-author Teichman joined the team after v1. Here's what I'll be using: David S Clausen, Sarah V Teichman, & Amy D Willis. 2025+. "Estimating Ratios of Means of Multicategory Data Observed with Sample and Category Perturbations." Biometrika, In Press. arxiv.org/abs/2402.05231.arxiv.orgEstimating Fold Changes from Partially Observed Outcomes with Applications in Microbial MetagenomicsWe consider the problem of estimating fold-changes in the expected value of a multivariate outcome observed with unknown sample-specific and category-specific perturbations. This challenge arises in h... 020
Amy D Willis @amydwillis.bsky.social · 16/12/2025Documentation: statdivlab.github.io/radEmu/ Blog post: statdivlab.github.io/blog/article... 4/4statdivlab.github.ioUsing Relative Abundance Data to Estimate of Multiplicative Differences in Mean Absolute AbundanceA differential abundance method for the analysis of microbiome data. radEmu estimates fold-differences in the abundance of taxa across samples relative to "typical" fold-differences. Notably, it does ... 030
Amy D Willis @amydwillis.bsky.social · 16/12/2025Feel guilty about replacing your zeroes with pseudocounts? Fed up with checking robustness to said pseudocount? Not sure what it means biologically to have a difference of 0.34 in averages of CLR-transformed data?* radEmu 🕶️🦤 3/ *Spoiler: It means nothing. It's inherently not interpretable. 120
Amy D Willis @amydwillis.bsky.social · 16/12/2025Oh yeah, ICYMI... radEmu is a method for differential abundance (or differential expression). It's especially well-suited to genomics and microbiome studies. 120
Amy D Willis @amydwillis.bsky.social · 16/12/2025Also, we made some common cases superdooper fast last week, so pull changes ⚡🏎️ Happy holidays, y'all! ❤️🦤❤️🦤❤️🦤❤️🦤❤️🦤❤️🦤❤️ 2/ 110
Amy D Willis @amydwillis.bsky.social · 16/12/2025I am delighted to share that the radEmu manuscript was accepted today to Biometrika 🥳 Sending gratitude to our users and reviewers for their enthusiasm and support!! Huge congrats to @davidandacat.bsky.social and Sarah Teichman on all their hard work and brilliance 😻❗ 1/ 2117
Amy D Willis @amydwillis.bsky.social · 08/12/2025With many thanks to @svteichman for bringing this to my attention and making these figures 010
Reposted by Amy D WillisAmy D Willis @amydwillis.bsky.social · 08/12/2025Thanks for corroborating!! We've seen again and again that ALDEx2's p-values can be either very conservative (= underpowered) or very anticonservative (= wrong), but this points to something more problematic, such as a mistake in how they're being calculated. 111
Amy D Willis @amydwillis.bsky.social · 08/12/2025Thanks for corroborating!! We've seen again and again that ALDEx2's p-values can be either very conservative (= underpowered) or very anticonservative (= wrong), but this points to something more problematic, such as a mistake in how they're being calculated. 111
Amy D Willis @amydwillis.bsky.social · 08/12/2025ALDEx2's p-values are unusually strongly correlated with its effect sizes. Has anyone else noticed this? That... shouldn't happen. Data is actual shotgun coverages, ~8000 "taxa", 57 samples. 172
Amy D Willis @amydwillis.bsky.social · 04/12/2025Let's do better science together by using agreed-upon definitions! Woohoo! 4/4 000
Amy D Willis @amydwillis.bsky.social · 04/12/2025Same article but without the paywall (sorry, I didn't think they'd paywall it! duh Amy...) github.com/statdivlab/p... 3/github.com 120
Amy D Willis @amydwillis.bsky.social · 04/12/2025There's some more info on confounders in microbiome studies (and how they differ from precision variables) in this perspective we wrote last year: www.nature.com/articles/s41... 2/ 100
Amy D Willis @amydwillis.bsky.social · 04/12/2025I had this conversation again with someone today, so posting a screenshot of an old t**** for those who missed it the first time around -- or want a refresher! 1/n 120
Amy D Willis @amydwillis.bsky.social · 11/11/2025With much gratitude to everyone who supported the wonderful #STAMPS summer course at the MBL, I share that STAMPS will not be offered in 2026. 💔😿 2107
Amy D Willis @amydwillis.bsky.social · 06/11/2025pixi is awesome. That is all. @titus.idyll.org you're all over this, right? 020
Amy D Willis @amydwillis.bsky.social · 04/11/2025Are you looking for a #tenuretrack or #openrank job in #biostatistics or #statistics? @fredhutch.org is hiring! Seattle is a wonderful place to grow your career; the bio/stat & science scene here is world-class; and you will work w/ amazing UW Biostat students. Apply! apply.interfolio.com/176402apply.interfolio.com Apply - Interfolio {{$ctrl.$state.data.pageTitle}} - Apply - Interfolio 0410
Amy D Willis @amydwillis.bsky.social · 03/11/2025dude you can just text me with this stuff ❤️ % of microbiome (& changes) aren't identifiable from HTS, but stacked barplots are helpful for generating clues esp if big shifts data snooping has implications for error rate control. there isn't a hypothesis you're interested in? good luck & enjoy xx 060
Amy D Willis @amydwillis.bsky.social · 31/10/2025In other news, I just had the great pleasure of teaching at #EBAME10 in Brittany 🇫🇷🥐. For this workshop, I made a tutorial showing how to use radEmu with the taxonomic profiling output of #anvio 😍😻 Workflow available for everyone on our blog: statdivlab.github.io/blog/article... ❤️❤️ 120
Amy D Willis @amydwillis.bsky.social · 31/10/2025Happy Friday from the StatDivLab! We just made updates to radEmu to make it faster & more stable than ever. Thanks to all our users -- we're excited to keep supporting it. github.com/statdivlab/r... Interpretable and robust microbial differential abundance, woohoo!github.comGitHub - statdivlab/radEmuContribute to statdivlab/radEmu development by creating an account on GitHub. 130
Amy D Willis @amydwillis.bsky.social · 18/10/2025Thanks for your interest, @genomesevolve.bsky.social ! As a field, *on average*, I feel we have moved... 1. away from obsessing over alpha and beta diversity comparisons 2. towards comparisons that are less sensitive to rare/undetected species (than diversity) So 📈 🥳 010
Amy D Willis @amydwillis.bsky.social · 23/07/2025Wisdom from @titus.idyll.org on the final day of #STAMPS2025: "There's free as in beer, and there's free as in kittens. #Bioinformatics software is free as in kittens. You have to love and care for them or they... well... yeah." 050
Amy D Willis @amydwillis.bsky.social · 22/07/2025The question was "...*on the same sequencing run*?" 000
Amy D Willis @amydwillis.bsky.social · 19/07/2025Given that a Poisson regression with log link targets the same parameter as your NB regression, I'd be curious to see the coverage of robust Wald CI's. `rigr` wraps this, so does `raoBust`, so should be easy to add. (Sorry -- I'm at a workshop today or I'd do it myself) 010
Amy D Willis @amydwillis.bsky.social · 19/07/2025raoBust doesn't invert score tests @nlaroy.bsky.social, but it does implement (model-misspecification) robust score tests, which are amazing for inference. Feel free to open a feature request. We'll see what we can do. github.com/statdivlab/r...github.comGitHub - statdivlab/raoBust: Generalized Linear Models with robust and non-robust Wald and Rao (score) testsGeneralized Linear Models with robust and non-robust Wald and Rao (score) tests - statdivlab/raoBust 000
Amy D Willis @amydwillis.bsky.social · 07/07/2025The StatDivLab is getting excited for #STAMPS2025 and working on our lectures for 🤩 Stats Day 😻. We ran this think-pair-share activity last year and, wow, the ensuing discussion was *very* engaged 😹😅 Reach out if you'd like to attend! Woods Hole, MA, July 14-24 2025. #microbiome #dataanalysis 161
Amy D Willis @amydwillis.bsky.social · 01/07/2025Agreed that (eg) MAG assembly vs taxonomic estimation makes a huge difference to your answer. Also, please definitely read the Conflict of Interest statement for the shallow shotgun paper and note how much of its claims rely on bioinfomatic subsampling *and not actual shallow sequencing data* 010
Reposted by Amy D WillisProf. Isabelle Laforest-Lapointe @ilafores.bsky.social · 30/06/2025Best workshop you can ever follow 021
Amy D Willis @amydwillis.bsky.social · 30/06/2025That's wonderful to hear, Isabelle!! Thanks for sharing, too. Hope to see you there in 2026. 010
Amy D Willis @amydwillis.bsky.social · 29/06/2025We had a last minute cancellation for STAMPS -- a phenomenal course on microbiome data analysis. Woods Hole, MA, July 14-24 2025. If you or *anyone* you know is interested to attend, please email Titus &/or me... and we'll do what we can!!! Thx for sharing widely! 🤞❤️ 1710
Amy D Willis @amydwillis.bsky.social · 21/05/2025The StatDivLab is *fully* reliant on our NIGMS #MIRA R35 to bring you top-quality statistical methods for microbiome research. #NIH They are quietly taking MIRAs away from this year's applicants. It's sneaky and scary. Fight back! Call your reps and tell them to protect science! 031
Reposted by Amy D WillisEthan White @ethanwhite.weecology.org · 21/05/2025It only takes 2 minutes to help defend our amazing program officers (and many other federal employees) and allow them to keep serving all Americans instead of responding to the whims of a single person 01313