Structural Biology

Inferring protein structure and conformational ensembles from experimental measurements.

Proteins occupy ensembles of conformations, but structural experiments only measure projections, diffraction patterns, spectra, and sparse restraints of them. I treat structure determination as inference: a learned structure predictor such as AlphaFold supplies the prior, and the experimental data supplies the evidence. The result is an ensemble that the measurements actually support, rather than a single best-guess structure. This work spans crystallography, NMR, and cryo-EM, along with understanding where predictors fail.

Selected work

* Equally-contributing first author, † Equally-contributing senior author

2026

  1. maddipatla2026foldem.png
    Under review2026
    Fold’EM: Direct atomic structure inference from cryo-EM particles
    A. Maddipatla, M. Mäeots, M. Pegoraro, N. Dräger, R. Covino, S. Vedula†, M. Pacesa†, A. Bronstein†
    Under review
    arXiv
  2. maddipatla2026density.png
    bioRxiv2026
    Density-guided AlphaFold3 uncovers unmodelled conformations in β2-microglobulin
    A. Maddipatla, S. Vedula, A. Bronstein, A. Marx
    bioRxiv
  3. maddipatla2026inference.png
    ICML2026
    Inference-time optimization for experiment-grounded protein ensemble generation
    A. Maddipatla, A. Rzayev, M. Pegoraro, M. Pacesa, P. Schanda, A. Marx, S. Vedula†, A. Bronstein†
    Proc. International Conference on Machine Learning (ICML)
    HTML
  4. maddipatla2025experiment.png
    Nat. Biotech.2026
    Experiment-guided AlphaFold3 resolves measurement-consistent protein ensembles
    A. Maddipatla, N. Sellam, M. Bojan, V. Masalitin, S. Vedula†, P. Schanda†, A. Marx†, A. Bronstein†
    Nature Biotechnology
  5. bojan2026representing.png
    ICLR2026
    Representing local protein environments with machine learning force fields
    S. Vedula*, M. Bojan*, A. Maddipatla, N. Sellam, A. Rzayev, F. Napoli, P. Schanda, A. Bronstein
    The Fourteenth International Conference on Learning Representations (ICLR)

2025

  1. vedula2025improving.png
    CSBJ2025
    Improving prediction accuracy in chimeric proteins with windowed multiple sequence alignment
    S. Vedula, A. Bronstein, A. Marx
    Computational and Structural Biotechnology Journal
  2. maddipatla2025inverse.png
    ICML2025
    Inverse problems with experiment-guided AlphaFold
    A. Maddipatla*, N. Sellam*, M. Bojan, S. Vedula†, P. Schanda†, A. Marx†, A. Bronstein†
    Proc. International Conference on Machine Learning (ICML)

2024

  1. maddipatla2024generative.png
    NeurIPS W2024
    Generative modeling of protein ensembles guided by crystallographic electron densities
    A. Maddipatla*, N. Sellam*, S. Vedula, A. Marx, A. Bronstein
    Machine Learning for Structural Biology Workshop, NeurIPS
  2. rosenberg2024seeing.png
    bioRxiv2024
    Seeing Double: Molecular dynamics simulations reveal the stability of certain alternate protein conformations in crystal structures
    A. Rosenberg, S. Vedula, A. Bronstein, A. Marx
    bioRxiv

← All research areas