Xinqiang Ding

Xinqiang Ding

Postdoctoral Associate

Massachusetts Institute of Technology

Biography

Hello! Welcome to my homepage!

I am a postdoctoral associate in the Zhang Lab at MIT. My current research of interest includes

  • Applying deep learning methods, especially deep generative models, to problems in computational chemistry and biophysics, such as free energy calculation, enhanced sampling, coarse-graining, and extracting useful information from molecular dynamics simulations.
  • Understanding chromosome strucures and dynamics using both data-driven statistical modeling and physics-based modeling;
  • Understanding intrinsically disordered proteins.
  • Connecting deep generative models with non-equilibrium statistical mechanics.

During my PhD study, I was fortunate to work with Prof. Charles L. Brooks III. The focuses of my PhD research include developing theorectical/computational methods for protein-ligand docking and free energy calculation and applying deep generative models to study protein evolution, fitness landscape and stablity.

I love collaborating with both theoretical and experimental researchers, so feel free to contact me if you have a project for which my expertise might be helpful.

Interests

  • Protein-ligand Docking & Free Energy Calculation
  • Chromsome Structures and Dynamics
  • Protein Evolution & Fitness Landscape & Protein Design
  • Molecular Dynamcis & MCMC & Enhanced Sampling & Machine Learning
  • Variational Inference & Probabilistic Deep Generative Models

Education

  • PhD in Bioinformatics, 2018

    University of Michigan

  • BSc in Pharmaceutical Science and Applied Mathematics, 2012

    Peking University

Experience

 
 
 
 
 

Postdoctoral Associate

Zhang Lab at the Massachusetts Institute of Technology

Jun 2019 – Present Cambridge, MA
 
 
 
 
 

Postdoctoral Fellow

Freedman Lab at the University of Chicago

Oct 2018 – May 2019 Chicago, IL
Developed methods for learning deep generative models using annealed importance sampling.
 
 
 
 
 

Graduate Student

Brooks Lab at the University of Michigan

Sep 2013 – Sep 2018 Ann Arbor, MI
Developed both physics-based and machine learning algorithms for drug discovery and protein engineering.
 
 
 
 
 

Research Assistant

Lu Lab at Tsinghua University

Sep 2011 – Aug 2013 Beijing
Collaboratively developed a method to improve RNA secondary structure prediction by incorporating experiment data into free energy model using machine learning methods.

Awards

  • PHYS Young Investigator Award, American Chemical Society (2021)
  • Infinite Expansion Award, MIT (2020)
  • Data Science GPU Grant, Nvidia Corporation (2019)
  • Rackham Predoctoral Fellowship, University of Michigan (2018)
  • DAEWOONG Foundation Scholarship (2010)
  • WU.Si Scholarship, Peking University (2009)

Publications

Quickly discover relevant content by filtering publications.
(2021). Generalizing the Discrete Gibbs Sampler-Based λ-Dynamics Approach for Multisite Sampling of Many Ligands. J. Chem. Theory Comput.

DOI PDF

(2020). Learning Deep Generative Models with Annealed Importance Sampling. Third Workshop on Machine Learning and the Physical Sciences (NeurIPS 2020).

PDF Code Poster

(2020). Computing Absolute Free Energy with Deep Generative Models. J. Phys. Chem. B.

DOI PDF

(2020). Accelerated CDOCKER with GPUs, parallel simulated annealing and fast Fourier transforms. J. Chem. Theory Comput.

DOI PDF

(2019). Deciphering Protein Evolution and Fitness Landscapes with Latent Space Models. Nat. Commun.

DOI PDF Code

(2019). Fast Solver for Large Scale Multistate Bennett Acceptance Ratio Equations. J. Chem. Theory Comput.

DOI PDF Code

(2018). CDOCKER and λ-dynamics for Prospective Prediction in D3R Grand Challenge 2. J. Comput. Aided Mol. Des..

DOI PDF

(2017). Gibbs Sampler-based λ-dynamics and Rao–Blackwell Estimator for Alchemical Free Energy Calculation. J. Chem. Theory Comput.

DOI PDF

(2017). Mechanism of Vps4 Hexamer Function Revealed by Cryo-EM. Sci. Adv.

DOI PDF

(2015). Improved Prediction of RNA Secondary Structure by Integrating the Free Energy Model with Restraints Derived from Experimental Probing Data. Nucleic Acids Res.

DOI PDF

Software

FastMBAR

FastMBAR is a Python solver for large scale MBAR/UWHAM equations running on both CPUs and GPUs. It can compute relative free energies of a large number of thermodynamic states used in alchemical free energy calculations and calculating potential of mean force (PMF) in umbrella sampling and temperature/Hamiltonian replica exchange simulations.