Research
A migrating cancer cell has to coordinate energy, cytoskeleton and protein synthesis at its leading edge. My thesis asks how it does that; my computational work asks what the same cells look like across thousands of tumours.
The Mitochondria–Actin–Translation Hub
Protrusion-enabling mRNAs are translated where they are needed. I show a physical coupling of outer-mitochondrial-membrane-proximal ribosomes, VDAC1-dependent energy supply and actin dynamics at the cell front, and test it with CRISPR-Cas9, proximity ligation, compartment-resolved proteomics, drug perturbation (FCCP, VBIT-12, translation inhibitors) and live super-resolution imaging.
- Minisymposium talk, ASCB|EMBO Cell Bio 2025, Philadelphia
- Expression-parameterised mechanotransduction ODE for ultrasensitive mechanosensor states and compartmentalised YAP regulation (under review)
- VDAC1 functional states in cancer (review, under review); mitochondrial allocation in metastasis (invited review, Seminars in Cancer Biology)
How telomere length talks to the tumour microenvironment
At CSIR-IGIB we found that the shelterin protein TRF2 binds outside telomeres and controls IL1R1 and TERT in a telomere-length-sensitive way, linking telomere state to NF-κB signalling and immune cell infiltration. The IL1R1 mechanism led to a dCas9–TRF2 fusion construct for targeted silencing, now a US patent application. I continue the theme computationally: a replicated patient-specific component of tumour telomere length across two pan-cancer cohorts, and cell-type-resolved senescence and telomere-maintenance programmes in immune ageing.
- Mukherjee*, Dutta* et al., eLife 2024 (co-first author)
- Sengupta et al., eLife 2025
- Patent application US20240117359A1
From actin images to nematic fields
Actin in a cell is a liquid crystal: it has orientational order, and that order breaks at topological defects. I use TV-L1 variational optical flow, nematic order and defect analysis, Helmholtz–Hodge decomposition and single-particle tracking on live-cell, confocal and super-resolution data (dSTORM, PALM, STED, TIRF, FLIM), and build deep-learning pipelines and napari plugins so that others can do the same.
- Actin nematic organisation in breast cancer (with Antonio Virgilio Failla, under review)
- Temporal order separates generic persistence from cue-directed commitment in cell migration (bioRxiv, under review)
- Mechanical observables of cancer invasion (Frontiers in Cell and Developmental Biology, 2026)
- Reanalysis of the JUMP Cell Painting genome-scale image screen
Regulatory, spatial and immune structure in public data
Public data holds structure that no single study was designed to see. As sole author I integrate TCGA, DepMap, CPTAC, ENCODE eCLIP and spatial transcriptomics (Visium, Xenium, CosMx) to find it: chromosomal-instability fields that mark antigen-presentation-low, immune-cold neighbourhoods; a regulatory atlas of 6,983 GWAS risk variants; healthy-organ immune baselines that track tumour immune load; AP-1 motif architecture in olaparib-adaptive chromatin remodelling.
- Frontiers in Systems Biology, 2026 (two articles)
- Three preprints on bioRxiv and medRxiv, 2026
- Foundation models for single-cell data, supported by a Google TPU Research Cloud award (2026)
Methods
What I can do at the bench, at the microscope and at the terminal.
- Molecular and cell biology
- CRISPR-Cas9 editing and base editing, pooled CRISPR screening in vitro and in vivo, molecular cloning, lentiviral production and transduction, proximity ligation assay, immunofluorescence, Western blot, ChIP, CUT&Tag, DNA and RNA FISH, immuno-FISH, multiplexed IHC, ELISA, qPCR, protein purification, calcium imaging.
- Cells and models
- Patient-derived organoids, PDX and PDOX models, orthotopic syngeneic breast cancer models (4T1, E0771), intraductal injection, primary cancer and T cell culture, iPSC culture, THP1 differentiation to macrophages and dendritic cells, macrophage infiltration and migration assays, multicolour FACS (FACSAria III, BD Rhapsody). Neuroblastoma, glioblastoma and liver cancer systems alongside breast cancer.
- Microscopy
- Confocal and Airyscan, dSTORM, PALM, STED, TIRF, FLIM, live-cell and live 3D time-lapse, single-molecule imaging, PRIMO maskless photopatterning.
- Image analysis
- TV-L1 optical flow, nematic order and topological defect analysis, Helmholtz–Hodge decomposition, single-particle tracking, FIB-SEM data analysis, napari plugin development.
- Machine learning
- PyTorch (foundation models, VAEs, GNN/GAT, contrastive learning, mixture of experts), JAX/Flax, automated architecture search, XGBoost, Bayesian optimisation.
- Genomics
- scRNA-seq atlas construction (Scanpy, Seurat), spatial transcriptomics (Visium, Xenium, CosMx), inferCNV, TCGA/DepMap/CPTAC/ENCODE integration, ChIP-seq, ATAC-seq, CUT&Tag and RNA-seq analysis, NGS library preparation, gene regulatory networks.
- Proteomics and simulation
- Compartment- and time-resolved quantitative proteomics, PRIDE and CPTAC reanalysis, differential abundance and pathway interpretation. GROMACS multi-microsecond MD, MM-PBSA, AlphaFold3, AutoDock and Glide docking.
- Infrastructure
- HPC with SLURM on H100/A100 GPUs and TPUs, Python and R pipeline engineering.