Saori Sakaue

I’ll be at #ashg2025 ! Please let me know if you want to chat with me about research, my lab at UW saorisakaue.github.io etc etc! I’ll also be standing at the poster session on Friday 5075F at 2:30-4p if you can swing by to chat👋

New paper from my lab and @jshendure.bsky.social lab! Led by the brilliant @zukailiu.bsky.social and @cxqiu.bsky.social. We tackled how anterior and posterior progenitor cells cooperate to self-organize into an embryonic structure (termed AP-gastruloid). (1/n) www.biorxiv.org/content/10.1...

The project was super fun and unexpected scientific journey, expanding my curiosity outside of the nucleus and DNA in genetics studies😀 I really appreciate warm mentorship from
@soumya-boston.bsky.social and invaluable inputs from genetics community in Boston! Hope you enjoy reading the preprint😉

Back to the original problem🔍 We've found that >50% of colocalization of causal variants between eQTL and neuropsychiatric disorders were specific to either nuclear or cytosolic eQTL! Subcellular localization of RNA and eQTL matters in identifying disease GWAS mechanisms (11/n

We've conventionally assumed that multiple causal variants in eQTL or GWAS is typically LD independent and working on different biological mechanisms (eg. enhancer AND promoter), but in our cases they can be LD *dependent* and work in concert to affect the same mechanism of RNA stability (10/n

Such examples with different causal variants between nucleus (enhancer) and cytosol (transcript) showed that sometimes many variants in complete linkage in cytosolic eQTL create a risk haplotype in the transcript, possibly suggesting a novel concept in multiple-causal-variant fine-mapping.(9/n

.. which makes cytosolic eQTL variants more asymmetric, downstream-skewed relative to the TSS as they localize within transcribed regions where RNA chemical modification can happen and affect RNA stability in the cytosol. (8/n

Surprisingly, 33% of eGenes had distinct causal variants between nucleus and cytosol for the same gene!
Nuclear early RNA was preferentially regulated by distal enhancers at the DNA transcription level, while cytosolic late RNA was regulated by variants within transcripts.. (7/n

More specifically, we asked if nuclear and cellular eQTL share the same causal variant (1), and if they are different, which genomic annotation(s) has preferential localization of causal variants for nuclear or cellular eQTL (2)? (6/n

We analyzed both nuclear and cellular (mostly cytosolic) RNA compartments and associated their RNA abundance with genotype in the brain and the kidney to achieve this goal! (5/n

So we asked distinct genetic regulatory mechanisms across entire RNA lifecycle by comparing eQTL between early RNA in the nucleus and late post-transcriptionally modified RNA in the cytosol. These molecular understanding will help us understand disease alleles precisely (4/n

But the DNA transcription into RNA is just the very first step among long journey of RNA lifecycle. RNA undergoes many processing, first in the nucleus e.g. splicing and polyA, then in the cytosol e.g. chemical modifications and degradation essential for mature RNA abundance (3/n

Problem: eQTL reveals disease alleles' function on gene expression, while it's been so puzzling🧐 that most #GWAS alleles do not colocalize with #eQTL. The traditional wisdom in the field is that eQTL regulate DNA transcription in the nucleus by altering regulatory sequences (2/n

📣Excited to share my last postdoc paper with
@soumya-boston.bsky.social on eQTL mechanisms depending on where the RNA is in the cell! @broadinstitute.org @harvardmed.bsky.social
TL;DR:Early RNA eQTL variants in the nucleus and late RNA eQTL variants in the cytosol have distinct molecular mechanism🧵

Powerful and accurate case-control analysis of spatial molecular data with deep learning-defined tissue microniches As spatial molecular data grow in scope and resolution, there is a pressing need to identify key spatial structures associated with disease. Current approaches often rely on hand-crafted features such...

Excited to tell you about VIMA (Variational Inference-based Microniche Analysis), our deep learning + stats hybrid for case-control analysis of multi-sample spatial molecular datasets. 🧵👇 www.biorxiv.org/content/10.1... @soumya-boston.bsky.social @broadinstitute.org