We can observe extraordinary phenomena. We cannot yet read them, predict them, or engineer them at will. That is the gap BIA Lab exists to close.
Cells communicate across tissue. Organisms self-organise from a single egg. Tumours hijack signalling networks with remarkable precision. We can observe all of this, but we lack the computational abstractions to understand, predict, or engineer these systems at will.
Three revolutions have converged. None alone would have been sufficient. Together, they make a new kind of science possible. BIA's timing here is deliberate, not accidental.
Foundation models and self-supervised learning now extract biological structure from images at a scale impossible before. The bottleneck is no longer seeing. It is knowing what to look for.
Light-sheet microscopy, expansion microscopy, and live multiplexed imaging let us observe single cells in full 3D context, over time, at subcellular resolution. That is the scale where biological information processing actually happens.
Spatially-resolved transcriptomics and 3D multiplexed tissue profiling produce data with genuine biological structure: maps of how cells communicate, organise, and decide. The raw material now exists.
We have already built foundational tools used by researchers worldwide, covering 3D cell segmentation, surface parameterisation, collective motion analysis, and quantitative phenomics. These are the building blocks.
The tools exist. What we are now pursuing is the deeper question: what is the logic that connects them? What are the computational primitives of biological organisation (the equivalent of data structures, control flow, and information channels) that will let us read, and eventually write, living systems as fluently as we write software?
That is the transition from tool-building to theory-building. From describing what cells do to understanding what cells compute.
This is the target. The reason it is worth pursuing is precisely because it would matter enormously if it came true.
The software built here becomes a standard toolkit for biological representation learning, used in labs studying development, cancer, immunity, and synthetic biology. The abstractions become part of how computational biologists think about cells.
A shared computational language for cellular organisation exists, the way TCP/IP exists for networks. Designing a tissue response in simulation before testing in biology becomes a real workflow, not a thought experiment.
We move from a science that catalogues what cells do to one that understands what cells compute. Therapeutic design, developmental engineering, and synthetic biology become more like engineering and less like guesswork.
A new discipline exists at the intersection of information theory, dynamical systems, and cell biology. The people who built it are remembered the way early computer scientists are: the ones who found the right abstractions.
We are looking for founding members: researchers and engineers who want to work on this problem from the ground up.
See Open Positions