MSc Bioinformatics
Covenant University
Ogun State, Nigeria
About Emmanuel Uwidia
Hey, I'm Emmanuel!
I'm a biochemist who fell in love with computers and eventually found my place at the intersection of protein science and machine learning.
I use AI and computational methods to study proteins, particularly how they can be characterized, engineered, and assigned biological functions. I also enjoy building complete machine-learning systems, taking projects from raw data and experimentation to working applications.
This is where I share my research, projects, and notes on interesting developments in AI for biology, along with the occasional random thought I simply felt was worth writing down.
Skills
Python
Biopython
PyTorch
NumPy
pandas
scikit-learn
Matplotlib
Docker
SQL
Git
Linux
Projects
EIHR-PROT
EIHR-PROT is a protein function prediction model for Gene Ontology prediction. Users can enter a protein sequence and receive the top-k GO term predictions across all ontology aspects. The model extracts embeddings from the ESM-2 protein language model and processes them through a two-layer MLP prediction branch. This output is adaptively weighted by a gating mechanism conditioned on explicit homology signals before being fused with a homology predictor to obtain the final prediction scores.
You can read about the model's architecture, trade-offs, benchmark performance and limitations here.
