Sperm DNA methylation epimutation biomarker for paternal offspring autism
doi.org
doi.org
ASD is widely acknowledged to be a highly heterogeneous condition which is likely to have multiple independent causes, and which has unclear boundaries with other diagnoses (especially ADHD)
So a study which effectively treats ASD as a single undifferentiated condition, and has rather small sample sizes too, I think should be treated with a great deal of scepticism.
Have you read this paper: https://www.nature.com/articles/s41398-019-0631-2
They find that the four-way distinction between ASD, ADHD, OCD and "typically developing" has poor correspondence to the observed interaction between cortical morphology and behavioural symptoms.
If you take their clusters – https://www.nature.com/articles/s41398-019-0631-2/figures/3 – as "what is really going on in the brain", then the same "what is really going on in the brain" can produce all four of ASD, ADHD, OCD and typical development, at different proportions (clusters 1-5). Whereas, conversely, there is a subgroup of ASD (cluster 10) which doesn't overlap with ADHD or OCD or typicality.
I'd also recommend reading https://link.springer.com/article/10.1007/s40489-016-0085-x
[0] See https://doi.org/10.1002/aur.1832 in particular the authors' claim that "our review of research has led us to conclude that using the ASD diagnosis in basic and translational research cannot provide valid conclusions because the diagnosis lacks both biological and construct validity" and also "The ASD diagnosis lacks boundary construct validity..."
Think about it, if you have 26 people randomly sampled and you give half a pill and the other half a placebo and the first half dies moments after ingestion, how confident are you that the pill is poisonous?
If the effect size is small, you may fail to detect it with a small population. If the sampling is biased, then you’re going to have a problem even if you have a massive population studied.
Think of it like this. The aim is to show a statistically significant difference. With a large enough sample, you will eventually show significance (hence Fisher's famous anecdote of "get more data"), but the effect size may be trivial. Whereas, with a small sample, only non-trivial effect sizes will achieve significance, and therefore achieving significance with such a small sample tells you something about the nontrivial extent of the effect size.
So you could argue that the extent of methylation change was so large, that it achieved significance in a sample even as small as n=13!
If you have some cluster of traits that don't hurt you, then imho they're just traits, not a disorder.
However, the boundary between BAP and ASD is not fixed. Different clinicians draw the line at different places, and it is moving over time. And which side of the line one ends up on can be determined by external life experiences. A person with BAP who lives a life full of luck and supportive environments may never be disabled by their traits to the extent that an ASD diagnosis is warranted. Give the same person worse luck and a more adverse environment, and the same traits may become much more disabling, and they may end up with an ASD diagnosis as a result. So, fundamentally the same person, whether they had ASD was determined not by who they are but by what their environment is.
This is part of what annoys me about some of these popular lines that "autistic people" and "neurotypical people" are "wired differently". Are people with BAP "neurotypical" or not? And sometimes the label is determined, not by one's "wiring", but by stuff going on in the world outside one's head, by society and culture and family and friends and fate and fortune.
It doesn't appear that the op made this claim.
I think I understand your perspective. I think there are some ASD associated traits which may result in sort of a tradeoff for some STEM workers. E.g someone who has a difficulty forming interpersonal relationships (negative) may thus have more free time to pursue academics and outperform their more social peers (positive). This seems more like what the op was talking about being able to observe.
I have no idea, but there are some sterotypes that they are below average. Would be interesting to see studies on that.
Software folks are really good about not getting divorced. And they have money, so they could if they wanted to.
But you probably wanted something more conclusive.
Rich people still get happy when they make money, and sad when they lose it.
It's kind of crazy to think otherwise.
- that graph cuts off at a pretty low annual salary (200k). That's certainly a comfortable lifestyle, but that's still within the realm of what normal people on the high end of a good job make. Maybe the happiness cut off is further into the aristocracy
- even if the increase in happiness is logrithmic in the amount of money, and that contunues. Logs grow slowly, maybe at some point it effectively doesn't matter.
Even more relevant is that some things could be cliff disorders, where it is actually beneficial till it actually falls over some point.
Like Icarus flying, but takes him too close to the sun.
Being obsessive about details making things or being paranoid about what you eat are probably great things to start with, but overrides its own utility at some point.
The cliff question would seem to be a very personal question. Where does that cliff exist? What traits create a cliff? ( not really looking for answer )
I have two autists in my house. For them autism has created some substantial struggles when dealing with the world in general. However, their ( what I call super powers ) have allowed them great success in their given fields ( cs and research ).
Both see the world substantially different than their mom and I do. Their view is often, simple and more forgiving.
Autists have and continue to make great contributions to the neural typical world. My hope is that we are not on a road to eliminate these individuals from our world.[0]
[0]https://www.appliedbehavioranalysisprograms.com/historys-30-...
For example, there’s fastidious, which is a great to have in at least some members of a team, or to be able to draw on as a skill or trait for oneself; and then there’s the same taken to an extreme which manifests as disorders such as OCD, hoarding, etc.
Autism Quotient (AQ) scores are generally accepted as a measure of Broad Autism Phenotype (BAP), which is subclinical ASD. That study produces the (entirely expected) conclusion that BAP is more common in men than women, and more common in STEM occupations than non-STEM occupations.
I'm not aware of any studies linking software engineering specifically to BAP, as opposed to STEM occupations in general. I think the evidence on that point is (thus far) based on anecdote and clinical experience.
That should be "Diagnosis of ASD has increased tenfold..." because of a heap of factors including the medical criteria of diagnosis changing, a historical underdiagnosis of autism in girls and women, and improved public knowledge of the condition leading to more assessments and diagnoses.
EDIT:
Towards the end, it says
> The frequency of autism in the population has dramatically increased over tenfold the past several decades. This increase appears to be due in part to increased diagnosis efficiency from 1975 to the early 2000s, as well as greater public awareness of the disease [3]. The more recent increase in the last couple of decades suggests environmental factors, and exposures also have a critical role in autism prevalence.
It is kind of sad given the advances in sequencing and state of computing power that we cannot reliably diagnose this stuff with sequencing. I thought the promise of 23andme, and ilk was to gain large amounts of data and make progress on these problems (NIPT seems like a success though).
Yeah having chuck norris in your study is going to skew any set of results.
> Sperm samples were obtained from fathers that have children with or without autism, and the sperm then assessed for alterations in DNA methylation. A genome-wide analysis (> 90%) for differential DNA methylation regions (DMRs) was used to identify DMRs in the sperm of fathers (n = 13) with autistic children in comparison with those (n = 13) without ASD children. The 805 DMR genomic features such as chromosomal location, CpG density and length of the DMRs were characterized. Genes associated with the DMRs were identified and found to be linked to previously known ASD genes, as well as other neurobiology-related genes. The potential sperm DMR biomarkers/diagnostic was validated with blinded test sets (n = 8–10) of individuals with an approximately 90% accuracy.
This has to be a preliminary study right? I mean n=13...
Found a mutation in spermatozoa that can be used as a flag for higher probability of having autistic children. Correct?
It's used for many things, e.g. turning on/off specific genes in somatic cells, repairing DNA (when DNA is copied, the old strand is methilated; so you know how to fix any differences between old and new strand), and more recently, for measuring aging (Horvath epigenetic clock).
We could say that the string of letters is not modified but the letters in the DNA string are "in cursive" now, so when we read the text we understand a different thing.
This gets close "Epigenetics in Families: Covariance between Mother and Child Methylation Patterns" https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7913850/
Example: Clinics harvests 12 eggs from a layd, 1 egg needed for IVF, IVF is succesful (baby born), lady is now happy to donate the remaining 11 eggs to science.
The shots a woman has to give herself daily to stimulate follicle production are no joke. That would be a lot to put someone through purely for research. It’s not as simple as that. And it’s wishful thinking to expect any woman to do that.