,
13 tweets,
3 min read
Read on Twitter
O autism-is-caused-by-the-microbiome study!
How shall I nope thee?
Let me count the ways...
How shall I nope thee?
Let me count the ways...
Human Gut Microbiota from Autism Spectrum Disorder Promote Behavioral Symptoms in Mice cell.com/cell/fulltext/… 
I am... skeptical. First, there's no good reason to think that the microbiome has much of anything to do with causing autism, but, fair enough, this paper is designed to test that idea...
The authors aim to test it by performing fecal transplants from autism spectrum disorder (ASD) or typically developing (TD) subjects into germ-free mice, and testing their behaviors...
As part of their rationale they describe the microbiome as being "heritable", thus implying that it may explain some of the heritability of ASD. Of course it is NOT heritable in the genetics sense of the word...
Because that would imply it is *differentially* inherited depending on the genetic relatedness. The microbiome may be transmitted but it is NOT heritable.
Anyhoo, the experiment is to test the recipient mice to see if they exhibit any "autistic" behaviors. I work on animal models of psychiatric disorders and even I don't think any such behaviours exist.
I'm certainly not going to the bank with *marble burying* in my back pocket...
Here are the actual data showing behavioral differences... I know they have p-values attached to them, but they don't pass the eyeball test... Sure would have been nice to have them replicated.
Here's another set, supposedly showing that treatment with taurine (one of the metabolites supposedly altered between the two sets of animals) alters some of these behaviors in a different "mouse model of autism"
Am I being unfair, in not being blown away by these supposed effects?
There are some additional data in the paper looking in a massively exploratory way at changes in gene splicing in the brain and at differences in metabolites between the two sets of animals...
Given the very high number of variables being probed, it's not surprising some differences are found. No replication samples are tested, so my prior expectation would be that these are spurious, until proven otherwise
