Iain Mathieson
@mathiesoniain
Followers
4K
Following
2K
Media
138
Statuses
2K
Population genetics, ancient DNA, statistics.
Philadelphia, PA
Joined April 2012
RT @ShaiCarmi: The Silicon Valley people interviewed seem almost apathetic to the realities of life of anyone outside their bubble. Their….
0
2
0
". and what did that produce? The cuckoo clock.”.
Many seem to believe that governments without a clear political majority lead to instability, chaos and economic doom . Switzerland .
1
0
0
The market can stay irrational longer than you can stay solvent.
1
3
26
Two incredible papers. Forget about ML. Coalescent HMMs are and always will be king.
0
20
90
Very annoying that people are taking seriously the genetically inferred bottleneck that is almost certainly wrong.
3
8
29
RT @sarakmathieson: Our paper "Interpreting GANs to infer natural selection from genetic data" is out in Genetics! Thank you @petrelharp as….
0
11
0
Completely agree - I always find claims in papers about the number of "novel" variants to be silly. Every position in the genome is variable somewhere. Patterns of LD *are* population specific but of more limited interest.
The use of the phrase “ancestry-specific variant” is increasing, particularly to describe rare Single-nucleotide variants (SNVs). But these alleles are not ancestry-specific. They have not yet been found elsewhere, but they will be. 1/n.
1
2
23
Couldn't agree more! Nothing wrong with PCA.
It's a pity that All of Us used UMAP to visualize ancestry variation in their new marker paper, out today in Nature. The UMAP algorithm, by design, exaggerates the distinctiveness of the most frequent ancestries, a message that can be misinterpreted by the public.
3
7
49
Uh. missing genotypes "might appear" as 0 instead of .
PSA, since I've now seen problems arising from this: As of Sept 2023, GATK no longer makes missing genotypes explicit in VCFs (i.e. "./." as in the VCF spec). Missing genotypes are instead coded as "0/0" with a DP=0 format field. See attached for examples.
3
6
40
RT @pontus_skoglund: "Disease Evolution in the Human Past" - Join us March 21st 2024 @TheCrick for a half-day symposium in the Medicine at….
0
58
0
Perhaps not only psychological traits!.
Developmental noise is an overlooked contributor to innate variation in psychological traits
1
2
9
RT @yun_s_song: After lengthy anticipation, we finally got to compare PrimateAI-3D with our Cross-Protein Transfer (CPT) model https://t.co….
0
23
0
Trying to access some data from EGA, but the access policy is blank and the link to the data access committee brings up a page saying "Oops. the server made a poo poo".
0
1
10
Also does not have to be a single factor. Since most things will decay in the same way, the explanation can be a combination of many environmental (and, yes, maybe genetic) factors, but I don't think it's really possible to distinguish them.
1
0
7
Anything that changes randomly every generation would decay exponentially. Imagine with probability p you adopt one of your parents' religions and w.p. 1-p the predominant religion. Then correlation of religion decays as genetic relatedness (+same assortative mating fudge factor).
1
2
11
It does seem to show that that 1) A single explanatory factor must decay with relatedness in a certain way and 2) if the explanation is genetic then there must be a very high level of assortative mating or population structure.
3
0
9
This is a good explanation. In my opinion this is the major limitation of the paper.
On the Clark paper: correlation functions often decay over various distances (genetic, environmental, spatial, etc). Observing a correlation that varies over genetic relatedness is not evidence that the cause is genetic, since many other processes create correlations that decay.
2
5
22