I am a computational biologist working in Dr. Douglas A. Lauffenburger lab (2019 fall-Present) at MIT. My research interests lie in developing machine learning and deep learning approches to reveal hidden immunological mechanisms in heterogeneous host-pathogen interaction using multi-omics and quantitative imaging. My recent research projects are computational method developments on pathogen genome sequences and cellular signaling to discover features that best predict survival states of individuals with infectious diseases (for example, tuberculosis, HIV and Ebola).
I have been collaborating with Dr. Galit Alter (Ragon Institute of MGH, MIT and Harvard) on COVID-19 related projects including Multisystem Inflammatory Syndrome in Children (MIS-C), Common Human Coronaviruses cross-reactivity, and Convalescent Plasma (CCP). I focus on using system serology to identify humoral immune response that correlates with disparate clinical phenotypes, which could provide critical insights into COVID-19 pathogenesis and therapeutics.
During my PhD, I worked with Dr. Kwonmoo Lee (now at Boston Children’s Hospital) on developing deep learning and machine learning models for quantitative cellular imaging in cytoskeleton dynamics. I dedicated on time-series data modeling to deconvolute subcellular protrusion heterogeneity and the underlying actin regulator dynamics from live cell imaging. Before that, as a research associate, I enhanced myself on Bayesian Statistical Learning and Global Optimization under supervision of Dr. Patrick Flaherty.
Download my resumé.
PhD in Biomedical Engineering, 2019
Worcester Polytechnic Institute (WPI), MA
MEng in Electronics Engineering and Computer Science, 2012
Peking University, China
BSc in Computer Science, 2009
Jilin University, China
We develop a deep learning-based framework “DeepHACKS” for subcellular phenotyping of protrusion dynamics to reveal fine differential drug responses at subcellular and single-Cell Levels.
Using Systems Serology, here we observed in 25 children with acute mild COVID a functional phagocyte and complement activating IgG response to SARS-CoV-2, comparable to the acute responses generated in adults with mild disease.
We highlight distinct humoral trajectories associated with resolution of SARS-CoV-2 infection and the need for early functional humoral immunity.
We establish a computational framework called HACKS (deconvolution of heterogeneous activity in coordination of cytoskeleton at the subcellular level) to deconvolve the subcellular heterogeneity of lamellipodial protrusion from live cell imaging.
We describe a deep transfer learning (DTL) based approach to process LDIH images in the context of cellular analyses.
We propose a novel framework called vU-net to reconstruct cell edge with a higher accuracy using limited training images.
By combining Canny edge detector, Convolutional Neural Networks (CNNs), and local intensity thresholding, we were able to detect cryptic lamellipodial edges of submarginal cells with high accuracy from the fluorescence time-lapse movies.