This is an odd blog post. I'm not sure the authors really understand the genetics they are trying to invoke. For example, when they say: > Research has shown t

This is an odd blog post. I'm not sure the authors really understand the genetics they are trying to invoke. For example, when they say: > Research has shown that some key autism genes are part of a shared ape heritage that predates the “split” that led us along a “human” path. This was when our ancient ape ancestors separated from other apes that are alive today. Other autism genes are more recent in evolutionary terms—though they are still more than 100,000 years old. This is a really misleading use of 'autism gene'. All this means is that a particular 'gene', in the sense of a large chunk of DNA which codes for a protein and not the usual sense of a distinct variant/allele present in only part of the population, when damaged or modified leads to autism. (So you would take all the GWAS hits and note that the SNPs tend to cluster in particular gene-regions.) This might tell you something about what biochemical or developmental pathways lead to autism when broken, but it doesn't tell you anything like 'autism is good' or 'autism helps humanity evolve'. At best, it tells you that autism is sometimes connected to relatively recent biological changes, which is not a surprise (what would an autistic chimpanzee look like?). You could make the exact same argument about schizophrenia: 'genes' (in the sense of large genetic units) in which schizophrenia SNPs often turn up are also often relatively recently evolved or related to the nervous system. Did 'humanity need schizophrenia to evolve'? Seems unlikely. The parsimonious interpretation is that the human central nervous system and other things have been evolving over the past 200k+ years and diverging from the other primates, supporting modern human intelligence and cognition, and autism and schizophrenia are simply ways that our brains can break and the newly-evolved stuff can break often thanks both to the constant introduction of new mutations and demographic events like Neanderthal introgression. (Neanderthal variants typically are harmful and are being evolved away. That must have come as a shock to those people claiming Neanderthal contributions made you smarter or something.) If you wanted to show that autism and genetic variants disposing to autism really does offer adaptive value or was important to the process, rather than simply reflecting the fragility of the human mind and long-standing mutation load, you would look for signs of selective sweeps for autism variants (either the rare variants of large effect or SNPs; none have been identified) or at least frequency-dependent selection (inconsistent with the additive genetic architecture), beneficial genetic correlations (there are none except some with intelligence), predicting any kind of better health (the opposite, indeed, just today: http://biorxiv.org/content/early/2017/06/09/148247 http://biorxiv.org/content/early/2017/06/09/148247 ); higher reproductive fitness in past or present populations (hah!) etc. The other arguments like 'cave art', well... I'll leave that to others.