Population genetics, evolution, human genetics, complex traits, functional genomics; also running, wildlife. Based at Stanford.
@JacobPhD
> what places would make the best US national parks if they hadn't been irrecoverably altered first?
That's a great question! Love your suggestion of Niagara NP.
When I lived in Chicago, I loved the tiny remnant of cypress swamp in southern Illinois (Cache River). Somewhere like this could have been the jewel of a midwest ecosystem park.
Ten more days to submit an abstract to #probgen23, held at CSHL on March 8-11th. Talks by Richard Durbin, Kathryn Roeder, 12 co-chairs, and chosen from abstracts. See the website for information on virtual attendance & financial aid: https://meetings.cshl.edu/meetings.aspx?meet=PROBGEN&year=23
Pondering: what places would make the best US national parks if they hadn't been irrecoverably altered first? E.g. Hudson Fjord National Park, Niagara Falls National Park, Hetch Hetchy National Park...
@gcbias sounds lovely! Happy birding for 2023!
RT @TwistBioscience@twitter.com
“#DNA is the world’s original storage medium- we’re just working to ⬇️ cost so it can be the future’s storage medium.” - @EmilyLeproust@twitter.com
@Chris_Mellor@twitter.com & @BlocksandFiles@twitter.com #DNADataStorage
Tape (and DNA?) needed to meet archive demand by 2030 https://bit.ly/3GbCXfB
🐦🔗: https://twitter.com/TwistBioscience/status/1609976572919808002
Happy New Year 2023!
What was your first bird of the year? (or other wildlife?)
Mine was Steller's jay followed by spotted towhee a moment later.
Adding more details. @anthony_herzig@twitter.com et al explore under which conditions adding diversity to the reference panel improves imputation accuracy.
They propose a cool trick to address the question. ->
RT @ShaiCarmi@twitter.com
Opening the Black-box of Imputation Software to Study the Impact of Reference Panel Composition on Performance
https://www.preprints.org/manuscript/202212.0483/v1
🐦🔗: https://twitter.com/ShaiCarmi/status/1608431603893800960
Always struck by how much more the House Sparrow body size maps match the weather on the worst days of winter than it does the average climate of a region. 1) Today's temperature map 2) House sparrow body sizes 3) USDA planting zones
RT @anshulkundaje@twitter.com
We published a paper
"Integrative single-cell analysis of cardiogenesis identifies developmental trajectories and non-coding mutations in congenital heart disease" 1/
https://authors.elsevier.com/c/1gIGbL7PXiquv
DOI: https://doi.org/10.1016/j.cell.2022.11.028
🐦🔗: https://twitter.com/anshulkundaje/status/1606009054727323648
@akundaje @Carldeboer @andyfraser
I think there is a conflation of multiple discussions here.
(1) The data are clear that complex trait variation is heavily shaped by selection*.
BUT it's usually difficult to measure this well at the level of individual variants, and the variants we detect are not the most strongly selected. So in this way I do not believe that selection is a great filter for causal variants.
*eg https://www.biorxiv.org/content/10.1101/2022.10.04.509926v1 among many others
Understanding selection is inherent to questions about *function*. Does a sequence have some biochemical activity because it has evolved to have that function, or because of mutational bias, chance, etc?
Whether or not you think this matters depends on what philosophical view you take about "function".
(As a distinct matter, selection plays a huge role in shaping the genetic architecture of traits, so it's relevant there, beyond considerations of philosophy)
@gcbias thanks Graham!
@alexwild Alex re Trump tax return release: One thing that becomes clear from the NYT reporting is how woefully outmatched the IRS was in dealing with this guy.
I think the release is useful, not to persuade the unpersuadable, but as a case study in how people with "complicated" finances and aggressive lawyers can defraud the government for decades and get away with it.
I have to assume there are many people doing this, and it's valuable to put this out in the open.
Fitness in this context refers to the various types of effects that selection can see, including viability and reproduction.
However, there's accumulating evidence that variants associated with late onset diseases are also under selection -- presumably because they have pleiotropic effects at earlier life-stages.
In general, it seems that when we map variants for complex trait X, selection is usually not mainly due to the fitness cost of trait X itself.
@andyfraser @akundaje
Agreed, in a protein coding sequence we have to make a distinction between *what* a nucleotide is versus *whether it is present*
For noncoding sequences we could think about nucleotides that sit between enhancers. The sequence may not matter, but the presence of those nucleotides may be important. And the effect of deleting one bp may be negligible, but the effect of deleting 1000 bp may be huge.
@akundaje
And if we think about function in terms of selection, there's some maximal total amount of function in the genome that an organism can maintain, depending on mutation rate and population size.
And the distribution of fitness effects across nucleotides must follow a power-law distribution, where some mutations have fitness costs of ~1 (eg nonsense mutation in an essential gene), and others are infinitesimally small (perhaps affecting GC content or nucleosome positioning)
@akundaje When we come to nucleotides in the genome, it seems better defined to me to think about function in terms of selection coefficients.
What would be the fitness cost of deleting/changing any particular nucleotide (or set of nucleotides)?
@akundaje
If dirt lodges in a pipe and thereby affects flow, do we say that the dirt has function?
(my instinct would be to say that the sound made by beating of the heart is not a function, but the dirt lodged in the pipe is... but I'm not a philosopher)
@akundaje There are a bunch of classic thought experiments about biological function.
I might say that the function of the heart is to pump blood... but then you say that it also makes sounds which a doctor can use to detect disease. Is making sound a *function* of the heart?
