Interested in knowing more about distance-based #phylogenetics? Check out #Structome Playground — a no-code interactive educational resource for exploring distance matrices, neighbour-joining trees, and now the resolution limit in real time.
biosig.lab.uq.edu.au/structome_pl...
#StrPhy (3/3)
Distance-based #StructuralPhylogenetics is a powerful approach for exploring deep evolutionary relationships. Tools like those in the #Structome suite make this accessible in web workflows — but there's always been a critical missing piece: confidence estimates for the inferred trees.
#StrPhy (1/3)
Like with the MBE work, DeepRoots introduces a new and much faster way to bootstrap. It achieves this by introducing "Embedding Jitter"—injecting Gaussian noise into the latent space to estimate branch support.
#Phylogenetics #StructuralBiology #Structome #Evolution #StructuralPhylogenetics #StrPhy
Protein structure is more conserved than sequence. Work in 2020 (MBE: academic.oup.com/mbe/article/...), showed 3D data unlocks deeper evolutionary signals. A thought was why not go higher than 3D? 🧵
#Phylogenetics #StructuralBiology #Structome #Evolution #StructuralPhylogenetics #StrPhy
Future work 1) inclusion of custom structure search 2) and I am working on digesting large scale text data via LLMs to offer meaningful text-based insights per query.
Feedback requested:
What would you improve?
#Structome #StructuralPhylogenetics #Science
Structome-TM preprint now live @ www.biorxiv.org/content/10.1...
@biorxivpreprint.bsky.social
#Science #Structome #StructuralPhylogenetics
#Structome began as an initiative to explore #evolutionary signals from #proteins sharing a structural neighborhood. It uses Q-score to rapidly gather datasets & build NJ trees. To highlight usage of Q-score, the tool is now called Structome-Q. App available here: biosig.lab.uq.edu.au/structome_q/