Jeff Spence
@jeffspence.github.io
assistant professor at ucsf interested in genetics, statistics, etc…
jeffspence.github.io
jeffspence.github.io
Pinned
Jeff Spence
@jeffspence.github.io
· Nov 7
Specificity, length and luck drive gene rankings in association studies - Nature
Genetic association tests prioritize candidate genes based on different criteria.
www.nature.com
How do GWAS and rare variant burden tests rank gene signals?
In new work @nature.com with @hakha.bsky.social, @jkpritch.bsky.social, and our wonderful coauthors we find that the key factors are what we call Specificity, Length, and Luck!
🧬🧪🧵
www.nature.com/articles/s41...
In new work @nature.com with @hakha.bsky.social, @jkpritch.bsky.social, and our wonderful coauthors we find that the key factors are what we call Specificity, Length, and Luck!
🧬🧪🧵
www.nature.com/articles/s41...
Reposted by Jeff Spence
Excited to see this out www.nature.com/articles/s41...! Nonparametric kernel-based tests for spatially variable isoform usage in spatial transcriptomics. So many interesting examples in the CNS and cancer, we're only scratching the surface!
Mapping isoforms and regulatory mechanisms from spatial transcriptomics data with SPLISOSM - Nature Biotechnology
Differential isoform usage is identified with high statistical power from spatial transcriptomics data.
www.nature.com
January 6, 2026 at 7:12 PM
Excited to see this out www.nature.com/articles/s41...! Nonparametric kernel-based tests for spatially variable isoform usage in spatial transcriptomics. So many interesting examples in the CNS and cancer, we're only scratching the surface!
Reposted by Jeff Spence
Clever use of proteomic data to stress-test TWAS and QTL colocalization methods, revealing a high false sign rate. This hypothesis about high-LD and cross-tissue confounding is particularly interesting:
January 6, 2026 at 5:52 PM
Clever use of proteomic data to stress-test TWAS and QTL colocalization methods, revealing a high false sign rate. This hypothesis about high-LD and cross-tissue confounding is particularly interesting:
Reposted by Jeff Spence
Happy to share our new preprint from @sashagusevposts.bsky.social and @nmancuso.bsky.social labs! We introduce Mr. PEG, a framework integrating perturbational screens, eQTL, and GWAS data to identify mediating genes for complex traits. (1/n) www.medrxiv.org/content/10.6...
Integrating perturbational screens, eQTL, and GWAS data identifies mediating genes for complex traits
Most current GWAS-eQTL approaches prioritize genes whose mediating effects on complex traits act through cis-regulation, while trans-acting genes remain largely underexplored. Recent perturbational sc...
www.medrxiv.org
January 5, 2026 at 10:27 PM
Happy to share our new preprint from @sashagusevposts.bsky.social and @nmancuso.bsky.social labs! We introduce Mr. PEG, a framework integrating perturbational screens, eQTL, and GWAS data to identify mediating genes for complex traits. (1/n) www.medrxiv.org/content/10.6...
Reposted by Jeff Spence
“Integrating perturbational screens, eQTL, and GWAS data identifies mediating genes for complex traits”
Very nice @medrxivpreprint.bsky.social study by @zeyunlu.bsky.social @sashagusevposts.bsky.social & colleagues 🧪🧬
www.medrxiv.org/content/10.6...
Very nice @medrxivpreprint.bsky.social study by @zeyunlu.bsky.social @sashagusevposts.bsky.social & colleagues 🧪🧬
www.medrxiv.org/content/10.6...
Integrating perturbational screens, eQTL, and GWAS data identifies mediating genes for complex traits
Most current GWAS-eQTL approaches prioritize genes whose mediating effects on complex traits act through cis-regulation, while trans-acting genes remain largely underexplored. Recent perturbational sc...
www.medrxiv.org
January 6, 2026 at 7:24 AM
“Integrating perturbational screens, eQTL, and GWAS data identifies mediating genes for complex traits”
Very nice @medrxivpreprint.bsky.social study by @zeyunlu.bsky.social @sashagusevposts.bsky.social & colleagues 🧪🧬
www.medrxiv.org/content/10.6...
Very nice @medrxivpreprint.bsky.social study by @zeyunlu.bsky.social @sashagusevposts.bsky.social & colleagues 🧪🧬
www.medrxiv.org/content/10.6...
Reposted by Jeff Spence
How well does TWAS estimate a gene’s direction of effect on a trait? We think of this as an important stress-test for the accuracy of TWAS.
In a new pre-print, we find that TWAS gets the sign wrong around 20-30% of the time!
doi.org/10.64898/202...
1/n
In a new pre-print, we find that TWAS gets the sign wrong around 20-30% of the time!
doi.org/10.64898/202...
1/n
High false sign rates in transcriptome-wide association studies
Transcriptome-wide association studies (TWAS) are widely used to identify genes involved in complex traits and to infer the direction of gene effects on traits. However, despite their popularity, it r...
doi.org
January 6, 2026 at 2:30 AM
How well does TWAS estimate a gene’s direction of effect on a trait? We think of this as an important stress-test for the accuracy of TWAS.
In a new pre-print, we find that TWAS gets the sign wrong around 20-30% of the time!
doi.org/10.64898/202...
1/n
In a new pre-print, we find that TWAS gets the sign wrong around 20-30% of the time!
doi.org/10.64898/202...
1/n
Reposted by Jeff Spence
New preprint alert: we use sign errors as a test of how well TWAS works.
Very worryingly we find that TWAS gets the sign wrong around 1/3 of the time (compared to 50% for pure guessing). You can read more about our analysis here, and what we think is going on 👇
Very worryingly we find that TWAS gets the sign wrong around 1/3 of the time (compared to 50% for pure guessing). You can read more about our analysis here, and what we think is going on 👇
How well does TWAS estimate a gene’s direction of effect on a trait? We think of this as an important stress-test for the accuracy of TWAS.
In a new pre-print, we find that TWAS gets the sign wrong around 20-30% of the time!
doi.org/10.64898/202...
1/n
In a new pre-print, we find that TWAS gets the sign wrong around 20-30% of the time!
doi.org/10.64898/202...
1/n
High false sign rates in transcriptome-wide association studies
Transcriptome-wide association studies (TWAS) are widely used to identify genes involved in complex traits and to infer the direction of gene effects on traits. However, despite their popularity, it r...
doi.org
January 6, 2026 at 2:48 AM
New preprint alert: we use sign errors as a test of how well TWAS works.
Very worryingly we find that TWAS gets the sign wrong around 1/3 of the time (compared to 50% for pure guessing). You can read more about our analysis here, and what we think is going on 👇
Very worryingly we find that TWAS gets the sign wrong around 1/3 of the time (compared to 50% for pure guessing). You can read more about our analysis here, and what we think is going on 👇
Reposted by Jeff Spence
Furthermore, Ron explored how perturbation outcomes grouped regulators that act coherently on transcriptional programs. Several known programs emerge, but the real magic is the context specificity: many programs and their regulators change drastically across different stimulation conditions.
January 5, 2026 at 6:42 PM
Furthermore, Ron explored how perturbation outcomes grouped regulators that act coherently on transcriptional programs. Several known programs emerge, but the real magic is the context specificity: many programs and their regulators change drastically across different stimulation conditions.
Reposted by Jeff Spence
Together with @ronghuizhu.bsky.social, we are thrilled to present our new perturb-seq study of 22M primary CD4+ T cells, across donors and timepoints – the result of a decade-long collaboration between the Marson @marsonlab.bsky.social and Pritchard @jkpritch.bsky.social labs 🧵 tinyurl.com/gwt2025
Genome-scale perturb-seq in primary human CD4+ T cells maps context-specific regulators of T cell programs and human immune traits
Gene regulatory networks encode the fundamental logic of cellular functions, but systematic network mapping remains challenging, especially in cell states relevant to human biology and disease. Here, ...
tinyurl.com
January 5, 2026 at 6:42 PM
Together with @ronghuizhu.bsky.social, we are thrilled to present our new perturb-seq study of 22M primary CD4+ T cells, across donors and timepoints – the result of a decade-long collaboration between the Marson @marsonlab.bsky.social and Pritchard @jkpritch.bsky.social labs 🧵 tinyurl.com/gwt2025
Reposted by Jeff Spence
Interesting to search for works with “Genetics” in the title in Anthropic Works List Lookup to see what was stolen and ingested.
#Genetics #AI #Anthropic
secure.anthropiccopyrightsettlement.com/lookup
#Genetics #AI #Anthropic
secure.anthropiccopyrightsettlement.com/lookup
Submit a Claim
secure.anthropiccopyrightsettlement.com
December 26, 2025 at 12:50 PM
Interesting to search for works with “Genetics” in the title in Anthropic Works List Lookup to see what was stolen and ingested.
#Genetics #AI #Anthropic
secure.anthropiccopyrightsettlement.com/lookup
#Genetics #AI #Anthropic
secure.anthropiccopyrightsettlement.com/lookup
Reposted by Jeff Spence
Congrats @jeffgroh.bsky.social et al. Some avocado trees open female-phase flowers in the morning & then male in afternoon. Others show complementary pattern (m->f), to synchronize pollination of two types. Jeff show this to be a >45Mya polymorphism at a transcription factor across 100s of species.
Balanced polymorphism in a floral transcription factor underlies an ancient rhythm of daily sex alternation in avocado https://www.biorxiv.org/content/10.64898/2025.12.22.695989v1
December 25, 2025 at 6:47 PM
Congrats @jeffgroh.bsky.social et al. Some avocado trees open female-phase flowers in the morning & then male in afternoon. Others show complementary pattern (m->f), to synchronize pollination of two types. Jeff show this to be a >45Mya polymorphism at a transcription factor across 100s of species.
Reposted by Jeff Spence
Boring, field-specific answer. This SMART-PTA technology, which allows genotype + phenotype in hundreds of somatic cells, will lead to an expansion of population genetics for somatic cells as a field. www.biorxiv.org/content/10.1...
December 25, 2025 at 4:28 AM
Boring, field-specific answer. This SMART-PTA technology, which allows genotype + phenotype in hundreds of somatic cells, will lead to an expansion of population genetics for somatic cells as a field. www.biorxiv.org/content/10.1...
Reposted by Jeff Spence
This is a really cool paper
In the December issue of #GENETICS, @jeremyjberg.bsky.social et al. introduce an evolutionary model of complex disease susceptibility, identifying how diseases are shaped by selection acting on other pleiotropically related traits. buff.ly/QTZPpYQ
December 22, 2025 at 11:04 PM
This is a really cool paper
Reposted by Jeff Spence
In the December issue of #GENETICS, @jeremyjberg.bsky.social et al. introduce an evolutionary model of complex disease susceptibility, identifying how diseases are shaped by selection acting on other pleiotropically related traits. buff.ly/QTZPpYQ
December 22, 2025 at 11:01 PM
In the December issue of #GENETICS, @jeremyjberg.bsky.social et al. introduce an evolutionary model of complex disease susceptibility, identifying how diseases are shaped by selection acting on other pleiotropically related traits. buff.ly/QTZPpYQ
Reposted by Jeff Spence
How much better is an ancestral recombination graph (ARG) than a site frequency spectrum (SFS)? For recovering mutation rate history, we can answer fairly precisely because both ARG and SFS are linear transforms of mutation rate history. This blog post uses spectral analysis to clarify the picture.
Observability of mutation rate histories from ancestral recombination graphs
This post explores mathematical aspects of recovering mutation rate histories from an ancestral recombination graph (ARG) Vs a sample frequency spectrum (SFS), expanding on a recent collaborative pape...
dewitt-lab.github.io
December 22, 2025 at 6:19 PM
How much better is an ancestral recombination graph (ARG) than a site frequency spectrum (SFS)? For recovering mutation rate history, we can answer fairly precisely because both ARG and SFS are linear transforms of mutation rate history. This blog post uses spectral analysis to clarify the picture.
Reposted by Jeff Spence
In fact, they appear to be eerily similar. The per generation mutation rate seems to lay between 10-9 and 10-8 per bp in all animal taxa surveyed to date–despite vast differences in environments, life histories, and three orders of magnitude variation in the generation time: 4/n
December 22, 2025 at 3:09 PM
In fact, they appear to be eerily similar. The per generation mutation rate seems to lay between 10-9 and 10-8 per bp in all animal taxa surveyed to date–despite vast differences in environments, life histories, and three orders of magnitude variation in the generation time: 4/n
Reposted by Jeff Spence
Happy to highlight an essay I wrote together with @marcdemanuel.bsky.social,
@natanaels.bsky.social and Anastasia Stolyarova, trying to think through what sets the mutation rate of a cell type in an animal species: www.biorxiv.org/content/10.6... 1/n
@natanaels.bsky.social and Anastasia Stolyarova, trying to think through what sets the mutation rate of a cell type in an animal species: www.biorxiv.org/content/10.6... 1/n
What sets the mutation rate of a cell type in an animal species?
Germline mutation rates per generation are strikingly similar across animals, despite vast differences in life histories. Analogously, in at least one somatic cell type, mutation rates at the end of l...
www.biorxiv.org
December 22, 2025 at 3:09 PM
Happy to highlight an essay I wrote together with @marcdemanuel.bsky.social,
@natanaels.bsky.social and Anastasia Stolyarova, trying to think through what sets the mutation rate of a cell type in an animal species: www.biorxiv.org/content/10.6... 1/n
@natanaels.bsky.social and Anastasia Stolyarova, trying to think through what sets the mutation rate of a cell type in an animal species: www.biorxiv.org/content/10.6... 1/n
Reposted by Jeff Spence
No contest. Just read the first two sentences of the abstract. www.nature.com/articles/s41...
All right it’s time for the annual “please tell us about one (or a few if you are ambitious) paper from 2025 that really impressed you and why we should all read it“! Go! If you tell us how it changed your view of the world and what makes it so powerful and consequential It would be excellent.
December 21, 2025 at 8:31 PM
No contest. Just read the first two sentences of the abstract. www.nature.com/articles/s41...
Reposted by Jeff Spence
This paper by Celentano et al. on scalable simulation under the birth death is one of my favourites for the year! It introduces some elegant thinning that has been sorely needed. Fantastic to pave the way forward for future simulation-heavy inference, including neural Bayes!
doi.org/10.1073/pnas...
doi.org/10.1073/pnas...
December 21, 2025 at 6:54 AM
This paper by Celentano et al. on scalable simulation under the birth death is one of my favourites for the year! It introduces some elegant thinning that has been sorely needed. Fantastic to pave the way forward for future simulation-heavy inference, including neural Bayes!
doi.org/10.1073/pnas...
doi.org/10.1073/pnas...
Reposted by Jeff Spence
I guess the preprint came out in 2024 but it was published this year so I'll say this paper from @jeffspence.github.io and @hakha.bsky.social which is probably the paper that pleiotropy-pilled me the most. Really got me to think about what GWAS means www.nature.com/articles/s41...
December 21, 2025 at 6:07 AM
I guess the preprint came out in 2024 but it was published this year so I'll say this paper from @jeffspence.github.io and @hakha.bsky.social which is probably the paper that pleiotropy-pilled me the most. Really got me to think about what GWAS means www.nature.com/articles/s41...
Reposted by Jeff Spence
All right it’s time for the annual “please tell us about one (or a few if you are ambitious) paper from 2025 that really impressed you and why we should all read it“! Go! If you tell us how it changed your view of the world and what makes it so powerful and consequential It would be excellent.
December 21, 2025 at 3:10 AM
All right it’s time for the annual “please tell us about one (or a few if you are ambitious) paper from 2025 that really impressed you and why we should all read it“! Go! If you tell us how it changed your view of the world and what makes it so powerful and consequential It would be excellent.
Reposted by Jeff Spence
High false sign rates in transcriptome-wide association studies https://www.biorxiv.org/content/10.64898/2025.12.19.695550v1
December 20, 2025 at 11:31 PM
High false sign rates in transcriptome-wide association studies https://www.biorxiv.org/content/10.64898/2025.12.19.695550v1
Reposted by Jeff Spence
But if each gene is doing multiple jobs, then every phenotypical trait is controlled by multiple genes.
Even homozygous deleterious mutations are doing something else when in heterozygous form: doi.org/10.64898/202...
Complexity is the norm, more complexity than what fits in the human imagination
Even homozygous deleterious mutations are doing something else when in heterozygous form: doi.org/10.64898/202...
Complexity is the norm, more complexity than what fits in the human imagination
December 19, 2025 at 5:29 PM
But if each gene is doing multiple jobs, then every phenotypical trait is controlled by multiple genes.
Even homozygous deleterious mutations are doing something else when in heterozygous form: doi.org/10.64898/202...
Complexity is the norm, more complexity than what fits in the human imagination
Even homozygous deleterious mutations are doing something else when in heterozygous form: doi.org/10.64898/202...
Complexity is the norm, more complexity than what fits in the human imagination
Reposted by Jeff Spence
This is a good time to talk about the TRUE genetics revolution brought in by sequencing the human genome:
The genetic underpinning of traits is not simple, will never be simple. Complex gene-gene interactions are the rule, not the exception 🧵
www.nature.com/articles/d41...
The genetic underpinning of traits is not simple, will never be simple. Complex gene-gene interactions are the rule, not the exception 🧵
www.nature.com/articles/d41...
Biobanks reveal genetic complexity in human evolution
Tiny genetic variations between humans, Neanderthals and Denisovans might not be all they were cracked up to be.
www.nature.com
December 19, 2025 at 5:29 PM
This is a good time to talk about the TRUE genetics revolution brought in by sequencing the human genome:
The genetic underpinning of traits is not simple, will never be simple. Complex gene-gene interactions are the rule, not the exception 🧵
www.nature.com/articles/d41...
The genetic underpinning of traits is not simple, will never be simple. Complex gene-gene interactions are the rule, not the exception 🧵
www.nature.com/articles/d41...
Reposted by Jeff Spence
Published online on Jan 2, 2025 and just appeared in the December 2025 issue!
A language model predicts the effects of genetic variants in the human genome go.nature.com/4gWppWg
rdcu.be/eUK6Z
rdcu.be/eUK6Z
A DNA language model based on multispecies alignment predicts the effects of genome-wide variants - Nature Biotechnology
A language model predicts the effects of genetic variants in the human genome.
go.nature.com
December 19, 2025 at 3:02 AM
Published online on Jan 2, 2025 and just appeared in the December 2025 issue!
A heads up for people using scipy.sparse.linalg.expm_multiply to solve linear ODEs -- the function appears to have some randomness that is not documented, but appears to be fixable by setting numpy.random.seed beforehand.
github.com/scipy/scipy/issues/24188
github.com/scipy/scipy/issues/24188
BUG: nondeterminism in scipy.sparse.linalg.expm_multiply · Issue #24188 · scipy/scipy
Describe your issue. I've tracked down an odd non-deterministic behavior in scipy.sparse.linalg.expm_multiply. Running the function on the same inputs multiple times can result in slightly differen...
github.com
December 18, 2025 at 11:23 PM
A heads up for people using scipy.sparse.linalg.expm_multiply to solve linear ODEs -- the function appears to have some randomness that is not documented, but appears to be fixable by setting numpy.random.seed beforehand.
github.com/scipy/scipy/issues/24188
github.com/scipy/scipy/issues/24188