Bacteriotherapy Design (using control theory principles) Postdoctoral Fellow

Opening for a Postdoctoral Research Fellow to join the Gibson Lab at Harvard Medical School (HMS), Brigham and Women’s Hospital (BWH), and the Broad Institute of MIT and Harvard. We leverage tools from machine learning (ML) and control theory to better understand biological systems and we are obsessed with uncertainty quantification. Control theoretic concepts are integrated both in the design of our optimization schemes and statistical machine learning models, as well as in the design of our in vitro and in vivo experiments. Our main area of focus is the microbiome and microbial dynamics more specifically with more recent work focussing on the gut-brain axis. Applications include the design of bacteriotherapies (bugs-as-drugs), developing methods to learn microbial dynamics at ecosystem-scale, studying the impact of phages on microbial communities, tracking low abundance pathogens, and methods for single cell biology. We focus on Bayesian methods that propagate measurement uncertainty so that we can assess confidence in model parameters and to help prioritize experiments. ML techniques applied include variational inference, Bayesian non-parametric models, discrete relaxation techniques (for making discrete models differentiable), and deep learning.

This specific project is focused on designing bacteriotherapies using control theory principles. Bacteriotherapies are defined communities of bacteria for use as a therapeutic, or simply stated “bugs-as-drugs”. We have already completed dense time series gnotobiotic studies with humanized (pre-print) and conventional mice (unpublished - ongoing NSF project), and

Using that data along with predictions from our dynamical systems inference models we will be

Using our dense time series gnotobiotic and Bayesian models we have generated ecological scale microbial interaction networks https://www.biorxiv.org/content/10.1101/2021.12.14.469105v2. Using predictions from those models we wish to design cocktails of bacteria that can manipulate a complex microbiome (for therapeutic applications). Simply adding the therapeutic taxa isn’t sufficient, however, as they may not robustly colonize the gut or may have large variations in carrying capacity depending on the background microbiome of the host. Using notions of robustness and stability from control theory we intend to augment bacterial cocktails with other microbial members to allow for more stable and robust colonization of the therapeutic taxa. One simple example could be the inclusion of a bacteria in negative feedback with the primary therapeutic taxa to regulate its abundance.

The candidate will not only design these bacterial communities but will help in the design of the mouse experiments, which would then be carried out by staff in the mouse facility. The candidate does not need to have experience with germ free studies and will not physically handle the mice. The candidate will however be in charge of all aspects of the experimental design. They will also likely contribute to the further development of our Bayesian models, expanding their capabilities, which may include the integration of multiple data modalities (e.g. spatial omics data) or exploring new inference techniques.

Qualifications

• PhD from a quantitative discipline (those with a pure bio background will be co-advised)
• Solid programming skills in Python; this isn’t a software engineering job, but you will need to be able to develop efficient implementations and apply your work to real data
• Strong publication track record
• Ability to reside in the U.S. and legally work in the country.

About the lab and fellow appointment

The Gibson Lab is physically located in the Division of Computational Pathology at BWH. All fellows have a triple appointment at BWH, HMS and the Broad. If necessary to conduct you work an official appointment at MIT can also be secured. For more information please go to the lab website: https://gibsonlab.io

Applications Process

Submit: (1) brief research statement (not to exceed 2 pages); (2) curriculum vitae; (3) two most relevant publications; (4) names and contact information of three individuals who can serve as references to: Travis Gibson, tegibson@bwh.harvard.edu. If you wish to chat briefly over Zoom before submitting materials to learn more details about our ongoing work, please inquire about this possibility.