Title:	Alpha Cell
DNr:	Berzelius-2025-61
Project Type:	LiU Berzelius
Principal Investigator:	Adil Mardinoglu <adilm@kth.se>
Affiliation:	Kungliga Tekniska högskolan
Duration:	2025-02-28 – 2025-09-01
Classification:	10610
Keywords:

Abstract

Alpha Cell is a large-scale computational biology initiative funded by the Knut and Alice Wallenberg Foundation (KAW). This project aims to develop AI-driven models capable of predicting cellular functions by integrating spatial-temporal molecular data. The project builds upon and extends the Human Protein Atlas (HPA), leveraging its comprehensive proteomic datasets to construct high-resolution 3D spatial maps of cellular structures and track their dynamic behavior over time. The primary objective of Alpha Cell is to simulate molecular interactions and cellular processes at an unprecedented scale using advanced computational modeling and machine learning techniques. By integrating proteomic, transcriptomic, and imaging data from HPA, this initiative will generate predictive models that elucidate intracellular processes, uncover disease mechanisms, and enable virtual testing of therapeutic interventions. The project also aims to systematically characterize unknown cellular states, contributing to a deeper understanding of human biology. To achieve these goals, Alpha Cell requires high-performance cloud computing infrastructure for large-scale data processing, AI model training, and simulation-based analysis. The computational demands include handling multi-terabyte datasets, executing deep learning models for spatial proteomics, and performing large-scale simulations of cellular systems. Secure and scalable cloud resources are essential for facilitating data storage, processing, and real-time collaboration among researchers. By leveraging cloud computing, Alpha Cell will enhance the accessibility and reproducibility of its AI-driven models, supporting the broader scientific community through open-access computational frameworks. The project represents a critical step in translating HPA data into predictive models of human biology, ultimately contributing to advancements in disease modeling, precision medicine, and therapeutic development.