Sadly, bc of ancient👴sample degradation, we've never explored the role of the 3D genome in archaic hominins like Neanderthals/Denisovans.
But lucky for us, new deep learning models can infer 3D genome structure from DNA sequence alone (like @gfudenberg & @drklly's Akita).
2/8
Using these models, we predict 3D contact maps across Neanderthal, Denisovan, and diverse modern human genomes.
We identify 167 distinct regions with diverged 3D genome organization between archaics and modern humans across the genome!
3/8
We show that 3D divergent windows in Neanderthals are enriched nearby genes associated with lots of interesting traits like the eye👀, supra-orbital ridges💀, hair💇, lungs 🫁, immune response💉, and cognition🧠! Provides a putative molecular mech for genotype -> phenotype
4/8
Despite the differences.. overall, modern human and archaic 3D genomes are more similar than expected, suggesting that the pressure to maintain 3D genome organization constrained hominin sequence evolution. (See the paper for some simulations & details.)
5/8
We also find that 3D genome organization constrained the landscape of archaic ancestry in humans today: regions more tolerant of 3D variation are enriched for introgression (archaic ancestry) in modern Eurasians.
6/8
Because we found 3D differences between humans and Neanderthals, we wondered...did humans inherit any of these novel predicted 3D genome structures from our Neanderthal ancestors bc of interbreeding? Spoiler alert: we did! 🤯
7/8
Lots more with more context + details in the paper. Plenty of limitations and future directions. We hope our deep-learning applications to predict 3D genome folding provides a window 🪟 into previously unobservable molecular mechanisms🧬. Thanks to all collaborators! Enjoy!
8/8
Artistic phylogeny inspired by Ahlquist, @FerVillanea@KWitt_aDNA et al. GBE 2021 Fig1 (not to scale)!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Reflecting on today's fantastic #iMED2022, I am thrilled to share an app I've been creating to create high-quality🤓, collaborative🗺️, & dynamic 📈diagnostic #MedEd schemas!
We have examples of how Neanderthals gave Eurasians individual genetic variants that contribute to traits; however, most medically/evolutionarily relevant traits are complex, with contributions from thousands of parts of the genome. We wanted something more comprehensive...
2/
We use partitioned heritability to investigate the relationship between introgression and diverse traits. First, we show that genomic regions with Neanderthal ancestry are depleted of heritability for all traits considered, except those related to skin and hair.
By synthesizing topologically associating domain (TAD) maps across 37 diverse cell types with 41 genome-wide association studies (GWASs), we investigate the differences in disease association and evolutionary pressure on variation across the 3D genome landscape.
2/n
We know that TAD boundary disruption by SV can lead to developmental disease and cancers. (Thus, selection acts against these SVs).
But what about the relationship between common human variation in TAD boundaries (eg. SNPs) and associations with complex traits?
3/n