How closely do you examine your stool 💩 🚽 ?

Colleagues from the #GermanPrimatCenter did just that and collected stool samples from #tupaias and measured the sizes of the contained particles! We studied the effect of age on food breakdown and indeed found that senile animals had signifcantly more very large particles in their stool.

So, make sure to brush your 🦷 and remember to chew your food properly.
(And beware of what a stool sample can tell...)

https://doi.org/10.1371/journal.pone.0274439

#MedStatGoe

❓ What methods are available for non-proportional hazards in clinical trials?
Our #SystematicReview is now available on arXiv
➡️https://arxiv.org/abs/2306.16858

Contributors include @bardo1 @stats_tobby@fediscience.or @poschm @friede1

#DataAnalysis #Statistics #MedStatGoe

@drdrago this is not too generic but very much with a specific software in mind 😉 , but I find the basic usage vignettes of the #rstats packages edgeR and DESeq2 excellent resources also for generic information.

edgeR: https://www.bioconductor.org/packages/release/bioc/vignettes/edgeR/inst/doc/edgeRUsersGuide.pdf

DESeq2:
http://bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html

#MedStatGoe

What is currently hot in Crosslinking and Proximity Labeling using #proteomics ?

If you are at the #GöttingenResearchCampus today you can find it out at this symposium of the #GPF:

'Proteins in company - approaching biomolecular interactions by mass spectrometry'

Excellent speaker lineup including Juri Rappsilber, Julien Béthune, and Ralf-Peter Jansen.

It is still OK to just drop by even when not pre-registered.

https://events.gwdg.de/event/438/

#MedStatGoe

⚠️ Be aware when you use xgboost in #rstats.

I was having trouble with constant performance measures when passing case weights (through #caret) in [0; 1]. Now, I am passing integer case weights and it seems to work fine.

#MedStatGoe

For the import of labelled SAS data into #rstats the function read_sas from the haven #rstats 📦 works just fine given a catalog file it likes.

#MedStatGoe

Don't export your data ❌ -- print it ✔️ !

At least if your data lives in SAS and you want the export to contain your painfully created value labels.

Oh, and be aware that SAS can't read its 'formats' file that contains the labels on another OS. 🙈

I'd be happy to hear about other ways to get labeled data from SAS into #rstats.

https://stackoverflow.com/a/62439325

#MedStatGoe

Ever wondered why you can't set MKL_NUM_THREADS from within #rstats ?

@henrikbengtsson has you covered (and introduces his 📦 startup):

https://henrikbengtsson.github.io/startup/#setting-environment-variables-during-startup

#MedStatGoe

Do you look at other domains 👓 ?

It was exciting to participate in the #KISSKI opening symposium yesterday. Especially nice to talk to colleagues from the #FraunhoferIEE working with #ai to model energy production and distribution. I learnt that graph neural networks are big there, too.

Interested in explaining decisions from graph neural networks? Have a look at our paper http://dx.doi.org/10.1186/s13073-021-00845-7

#MedStatGoe