Robert Hoehndorf
leechuck.bsky.social
Robert Hoehndorf
@leechuck.bsky.social
Associate Professor in Computer Science at KAUST. Editor in Chief for the Journal of Biomedical Semantics. Interested in knowledge representation, bioinformatics, neuro-symbolic AI.
We made three pangenome graphs 🧬 public, one for the Japanese, one for the Saudi population, and a merged graph (JaSaPaGe). Useful for 🖥️ bioinformatics on either population, or to evaluate how pangenome graphs behave when two different populations are included. jasapage.bio2vec.net/view for PanGene
Phased genome assemblies and pangenome graphs of human populations of Japan and Saudi Arabia
The selection of a reference sequence in genome analysis is critical, as it serves as the foundation for all downstream analyses. Recently, the pangenome graph has been proposed as a data model that i...
www.biorxiv.org
December 23, 2024 at 9:05 AM
Reposted by Robert Hoehndorf
ProtBoost: protein function prediction with Py-Boost and Graph Neural Networks -- CAFA5 top2 solution

The second-place solution in CAFA5 has now been published.
Paper: arxiv.org/abs/2412.045...

GitHub Repo: github.com/btbpanda/CAF...

Kaggle Writeup: www.kaggle.com/competitions...
CAFA 5 Protein Function Prediction
Predict the biological function of a protein
www.kaggle.com
December 17, 2024 at 10:04 PM
Proud to share our new paper! A complete genome from Saudi Arabia (KSA001), freely available to all. Complex work - not just sequencing & assembly challenges, but also navigating IRB approval to ensure ethical data sharing & open science principles.
nature.com/articles/s41...
A reference quality, fully annotated diploid genome from a Saudi individual - Scientific Data
Scientific Data - A reference quality, fully annotated diploid genome from a Saudi individual
nature.com
November 24, 2024 at 11:12 AM
I am excited to share that our paper on creating a very large structure causal model for diseases has been published. We generate an SCM containing most common diseases, and validate with #UKBiobank data, for better polygenic scores, and finding pleitropic variants academic.oup.com/bioinformati...
Causal relationships between diseases mined from the literature improve the use of polygenic risk scores
AbstractMotivation. Identifying causal relations between diseases allows for the study of shared pathways, biological mechanisms, and inter-disease risks.
academic.oup.com
November 18, 2024 at 4:26 AM