Research overview
We use interdisciplinary methodologies to investigate the cell biology, biophysics, and biochemistry of mitosis. Trainees obtain rigorous training encompassing molecular and cell biological methods, biochemistry, quantitative imaging, image analysis, and mathematical modeling.
AI-aided de novo protein design for engineering new cytoskeletal proteins
Rapid advances in AI-based structure prediction and generative protein design — including AlphaFold, RFdiffusion, and ProteinMPNN — have opened an entirely new frontier: designing proteins with prescribed folds and functions from scratch, without relying on natural templates.
We are harnessing these tools to engineer novel cytoskeletal proteins. By computationally designing proteins that can polymerize, form filaments, or interact with existing cytoskeletal elements in defined ways, we aim to build synthetic cytoskeletal systems with tunable mechanical and dynamic properties.
The cytoskeleton is the mechanical backbone of the cell. Engineering new cytoskeletal proteins will let us probe the design principles of filament assembly, test how mechanical properties arise from molecular structure, and ultimately build programmable intracellular machines.
Generative AI design
Using RFdiffusion, ProteinMPNN, and AlphaFold to design and validate novel protein folds
Protein characterization
Biochemical and biophysical validation of designed structures, assemblies, and interactions
In vivo testing
Expressing designed proteins in cells and assessing their cytoskeletal organization and function
In vitro reconstitution
Reconstituting filament assembly and dynamics with purified designed proteins
Perturbation and adaptation of the mitotic checkpoint in cancer biology
The mitotic checkpoint is a composite system of interconnected biochemical sub-systems: a mechanical switch that turns on and off; a signaling cascade that conveys the state of this switch; and a biochemical switch that drives the cell-cycle transition.
We investigate the systems biology of the signaling cascade to understand how natural variation in signaling protein expression levels affects checkpoint output. This question has direct bearing on cancer biology, where aberrant protein expression is widespread and checkpoint fidelity is frequently compromised.
To measure checkpoint behavior at scale, we use CRISPR-edited cell lines combined with high-throughput live-cell microscopy, and an in-house neural network (Virdi and Joglekar, MBoC) to automatically quantify signaling outputs across thousands of cells. These data feed quantitative mathematical models of the checkpoint.
Chromosome missegregation plays a fundamental role in tumorigenesis and the emergence of drug resistance in cancer cells. Understanding how cancer cells perturb and adapt to checkpoint dysfunction is essential for developing new therapeutic strategies.
CRISPR engineering
Endogenous tagging of checkpoint proteins in human cell lines without overexpression artifacts
eSAC system
Ectopic spindle assembly checkpoint activator — a patented synthetic biology tool for controlling checkpoint duration
Mathematical modeling
Quantitative ODE models of checkpoint signaling built from in vivo biochemical measurements
In vivo biochemistry
Measuring individual reaction rates and protein abundances directly in living cells
Key publications in this area:
Reverse engineering kinetochores using de novo-designed proteins
This is an exciting new frontier for the lab. Decades of investigations of the kinetochore have built a complete picture of its structure, function, and regulation. What comes next?
We are pioneering efforts to reverse engineer the kinetochore using de novo-designed proteins. The long-term goal of this research is to build simplified kinetochore-like machines. This project relies heavily on a combination of in vivo and in vitro investigations with applied protein design.
Ajit spent a 6-month sabbatical in David Baker's lab at the University of Washington, learning the basics of de novo protein design. He returned to Michigan with the vision to apply these tools to re-engineer the kinetochore from scratch.
Computational protein design
Design of novel protein folds and binding interfaces not found in nature
Protein characterization
Biochemical and biophysical validation of designed protein structures and interactions
In vivo testing
Expressing designed proteins in cells and testing their ability to perform kinetochore-like functions
This is an active and rapidly evolving research direction — watch this space for publications. Interested in joining?
Interdisciplinary training
Lab members gain broad, rigorous training that spans multiple disciplines — ideal for careers in academia, biotech, or medicine.
Molecular & cell biology
Cloning, CRISPR, cell culture, microscopy
Quantitative imaging
Fluorescence microscopy, image analysis, FRET
Biochemistry
Protein purification, in vitro reconstitution, binding assays
Computational methods
Mathematical modeling, data analysis, protein design