Hashem Koohy

I lead a multidisciplinary research group at Oxford, where we combine immunology, AI, systems biology, and structural modelling to decode how T cells interpret molecular information and shape adaptative cellular immunity. I also founded ImmSilico, a consultancy supporting researchers, biotech teams, and organisations in designing smarter experiments, integrating complex datasets, and applying modern AI tools to accelerate immunology discovery. I also host Unravelling T Cell Recognition a research-focused webinar series that brings together leading scientists to discuss seminal advances at the intersection of T cell biology and artificial intelligence. My mission is to bridge research, technology, and communication, ultimately advancing our understanding of adaptive immunity while empowering the next generation of scientist and innovators.

Machine learning and integrative analysis of high throughput sequencing data in cancer immunotherapy.

I have capitalised a collaboration between WSB mathematicians, Cardiff University Immunologists and Babraham Institute Immunologists to investigate the changes in diversity of antigen receptor repertoire over time (age) and between individuals (and replicates) from different conditions such as healthy vs infected or immunized.

As a member of the Nuclear Dynamics Programme (led by Dr. Peter Fraser), I was in charge of integrating the large genomics data into supervised and un- supervised machine learning techniques to further explore: a) the mechanisms underlying the diversity of the antigen receptor repertoire in mouse B lymphocytes, and b) the chromatin changes associated with age.

I was in charge of developing machine learning techniques for further understanding of transcriptional regulation. Specifically, I developed a multivariate Hidden Markov Model (named Composure) for genome-wide detection of regulatory regions in human genome. I also discovered the sequence specificity of DNase Hypersensitive enzyme and presented a Position Weight Matrix model for the removal of the resulting bias.

1- Developed an alignment-free (mathematical) model for comparison of cis-regulatory modules in D.melanogaster genome. 2- Implemented the model in C++ in a dynamic programming paradigm to tackle it very intense computation space. 3- Used model to predict the likely regulatory regions associated to fly olfactory systems and the most likely transcription factors.

Developed an alignment-free (mathematical) model to compare the functionality cis-regulatory modules. Implemented the model in C++ using a dynamic programming approach. I used the model to predict the regulatory regions associated to the D. Melanogaster olfactory systems as well as the underlying transcription factors.

An MSc degree in Mathematical biology: courses included: Mathematical Modelling, Statistics/Bioinformatics, Molecular Simulations, Numerical Methods, Data Acquisition. Projects included: 1- Simulation of spatio-temporal protein distribution with GDF (with Dr. Markus Kirkilionis). 2- Decoding calcium signals in human myometrium (with Dr. Anatoly Shmygol). 3- Measurement of DNA persistence length (with Prof. Alison Rodger).