There goes the kit-ome!

I loved reading it! My thoughts on this:

bsky.app/profile/fili...

Real wins so far: recurrent C. diff, maybe microbiota-directed foods in child malnutrition. That is basically it. Mitchell et al. are right: either we get power, clear mechanisms and honest effect sizes, or we admit this line is a dead end.

Things the microbiota was supposed to cure but did not: obesity, diabetes, autism, depression, IBS, cancer immunotherapy failure. FMT or probiotics did not deliver. The whole thing rested on the meaningless idea of “dysbiosis”.

I started in microbiome work around 2011. We massively oversold stool 16S. Every bar plot was treated as biology instead of compositional chatter driven by diet, site, pipelines and confounders we barely measured.

They go through human cohorts, mouse work, and clinical trials. Across all three: small samples, mushy hypotheses, bad stats, no robust replication, and tiny effect sizes when they exist at all.

Look at Fig. 3. The same taxa reported as ↑, ↓, or “no change” across autism–microbiome papers. If you can call all of that “dysbiosis”, you do not have a disease signature. You have noise with good PR.

Tu crois que ça implique des cigares?

Anything can be a CNS neuron if you are bold enough.

J’y touche pas, ça doit mordre.

Oui, la condition actuelle à Lyon est d’être inscrit•e au diplôme de l’ENSL.

Lyon 3 be like