Sasha Gusev
@SashaGusevPosts
Followers
20K
Following
21K
Media
3K
Statuses
20K
Statistical geneticist | Associate Prof at @DanaFarber / @harvardmed / @DFCIPopSci | Blogging at https://t.co/4D7UObBNdd
Joined January 2023
I've written about race, genetic ancestry, analyses of large biobanks, and human history https://t.co/DogeRfUCdr I'll summarize the key points here 🧵:
43
291
1K
Big new blogpost! My guide to data visualization, which includes a very long table of contents, tons of charts, and more. --> Why data visualization matters and how to make charts more effective, clear, transparent, and sometimes, beautiful. https://t.co/hDQhDL5rR1
14
160
1K
🧬 Do genetic variants interact in humans? For years, the answer was “probably, but we lacked solid statistical evidence.” We now bring one of the strongest pieces of evidence to date. Thrilled to share our latest work, FAME, now at Nature Genetics:
nature.com
Nature Genetics - FAME is a biobank-scale method that tests whether the effect of an SNP on a quantitative trait is modulated by an individual’s polygenic background. FAME can also be used to...
12
45
229
Immigration restrictionists don’t want to just own up to racist motives so they end up endorsing absurd economics.
The White House says that meritocracy should not be used as justification for recruiting foreign talent in its latest national security strategy document. https://t.co/f1iwGVMahf
57
61
518
D1 Powerhouses Trust INMOTION Air! Minnesota • Indiana • USF • NDSU • WSU • Missouri • UND • Purdue • Iowa • Georgetown • Dartmouth • Illinois • Northwestern • Nebraska • Idaho State Batting Cages, and Turtle Backstops powered by AIR Website: inmotionair
0
0
3
I rounded up some of the missing heritability discussions from the past few weeks and got the last word in (for now).
7
27
127
Please take a moment to ponder the ringing stupidity of the concept "real biological name"
HHS just updated Rachel Levine’s official portrait to his real biological name: Richard An HHS spokesperson said: "We remain committed to reversing harmful policies enacted by Levine and ensuring that biological reality guides our approach to public health."
115
202
5K
2016-2022: we really thort single cell and spatial would take genomics to that next level. By asking deeper questions and having the tools to answer them. That would in turn feed more funding and more research positions and vendor investments. Instead:
7
6
84
Our AI-powered platform helps you deliver digital innovation faster and with less risk by providing a fundamentally better approach to test automation.
59
104
608
Sydney Sweeney tells PEOPLE that she “doesn’t support the views” that people connected her American Eagle campaign to: “I was honestly surprised by the reaction. I did it because I love the jeans and love the brand. I don’t support the views some people chose to connect to the
5
6
192
Rewatching #BreakingBad in 2025 with fresh eyes as #Radonc resident☢️ Let's talk about how Walter White's Stage IIIA lung adenocarcinoma would be treated if diagnosed today instead of 2008. 🧵 #lcsm #MedTwitter Because honestly? The show may look very different. Here's why...
7
39
206
This is accurate and why the impact factor problem needs to be solved from the top down. Senior academics who lead search committees should pledge to review all applications blind to journal/IF and pressure other institutions to commit to the same.
I tried to tackle hard questions in my research, thinking people would notice and I’d eventually get a TT job. I had one zoom interview in 5 years of applying. The system works the way it works; now I tell people don’t make my mistake of thinking it works the way you wish it did.
4
5
62
3 questions to test your intition about genetics and CLT, inspired by @davidbessis argument that "long tailed phenotypes are not compatible with high additive h2": 1 Assume all traits are additive on the genetic level, no gxg, and no gxe exists. Is Bessis' argument correct? 2:
1
4
20
When Medicare payments to physicians fall behind the rate of inflation, everyone feels the strain. Learn how these cuts affect you.
0
2
6
The high correlation for DZ twins apart in this landmark study implies a massive selective placement effect and basically makes the data unusable. Perhaps that explains why the result went unpublished for 35 years.
Actually, Segal et al (2025) reported the DZA correlations in the MISTRA sample, albeit in a paper that was mostly about drawing tests. Comparing the MZA correlations to the DZA correlations yields an estimated FSIQ heritability of 24% and g-factor heritability of 52%. Of
5
4
83
I've yet to see an answer from twin study die-hards to this discrepancy. Ironically, the response feels very "nurturist" -- the methods are too complicated, there must be some confounding, they estimate "different" parameters (without stating which), don't talk about it.
1
1
26
Rare, moderately penetrant variants identifiable by whole-genome sequencing were one of the core hypotheses for "the case of the missing heritability" ( https://t.co/itH1nDu2uO) so it is useful to put that hypothesis to bed (on average).
2
3
23
I think the "hereditarian vs. nurturist" framing in the recent ACX post on heritability misses the most interesting result: that multiple different methods now show that twin study estimates are inflated for simple biomedical traits and the gap is not explained by rare variants.
However, if the pedigree-based analysis cannot be used as a stand-in for the classical twin design (i.e. it would imply lower equilibrium heritability estimates than the twin method in the same cohort), then the small 12% difference doesn't really tell us much about the gap
3
11
120
Mini ATG Gym Design + Knee & Lower Back Recipes That Have Been Wonderful For My Clients (Heads up: long article format) I measured off 200 square feet (10’ x 20’), including ample walking space, and polished off my own mini ATG. In fact, only 100 square feet of it is really
15
72
965
I agree with @slatestarcodex that the latest WGS analysis of UK Biobank participants has not resolved the missing (IQ) heritability debate for me. However, I think Scott's article makes a bizarre claim and misses some important arguments that I wanted to draw attention to.
substacktools.com
...
4
20
74
@SashaGusevPosts Have been urging the QTL/GWAS community to routinely profile caQTLs for years especially with smaller sample sizes. Hopefully this convinces folks to pursue the chromatin +expression route. H3k27ac and me3 QTLs would be even more informative. Just not as convenient at single cell
0
1
5
Excited to share our new FinnGen single-nucleus multiome preprint! 🧬 We profiled ~10M PBMCs (snRNA-seq + snATAC-seq) from 1,108 Finnish donors to map how genetic variants drive complex disease through chromatin and gene regulation 🧵👇 🔗 Link: https://t.co/A0LDn0jAKP
6
43
188
This is brutal. Jones is making "death recorded"-level errors.
Garett Jones has been critical of my work from a distance, but not in a formal way. More just gossipy comments to friends, colleagues, and acquaintances. I decided to respond when he put them in writing. Read them, you won’t regret it. https://t.co/IpNQk8J1tS
3
9
151
Consistent with recent findings by Rosen et al. ( https://t.co/lW89xLqXZF) that eQTLs detected at larger sample sizes (with more statistical power) converge on GWAS associations. The No. 1 rule in germline genetics is sample size is king!
pubmed.ncbi.nlm.nih.gov
Expression quantitative trait locus (eQTL) studies in human cohorts typically detect at least one regulatory signal per gene, and have been proposed as a way to explain mechanisms of genetic liabil...
0
4
10