Pressé Group

What We Do

At the single molecule level, experiments now follow the steps in the life's journey of a single protein; from its synthesis in a ribosome, to its activity in a complex dynamical environment, to its death by proteolysis. At the cellular level, experiments reveal with incredible detail how groups of proteins come together to regulate important events such as cell division. In an ideal world, experiments should not require much modeling to reveal physical insight -- the data should be self-evident. Yet biophysical data is noisy, complex and largely incomplete for a variety of reasons. This is especially true of data collected from live cells.

On the theory side, we develop, adapt and use tools derived from statistics, statistical physics and stochastic processes, broadly defined, to understand living systems across multiple time and length scales. On this front, there are two main research directions in our group: 1) we develop methods to infer models from imaging and spectroscopy data in biophysics with a recent focus on Bayesian nonparametrics; 2) we are developing models to understand enzymatic and molecular motor efficiency. On the experimental front, we are exploring the role of hydrodynamics on the interaction of bacterial predators with their prey.

See projects tab for a description of specific research directions.

Our Youtube Channel

Talks Online: Often times, the best way to communicate our science to a broad audience is to post scientific talks we've given. I will try to record and post some of my (scientifically non-overlapping) talks on my youtube channel under uploaded videos. For the moment, we just got started doing this.

Figures in the Slideshow

Proteins stepping on landmines: Enzymes diffuse faster after catalyzing chemical reactions. See our Nature paper for details. See our publications tab for links popular science articles written on this work. Credit for the figure: iSO-FORM,

Staying one step ahead: single molecule data collected by the Marqusee Lab at UC Berkeley. These single molecule data show extension of a single RNA hairpin as a function of time. The molecule undergoes zipping and unzipping transitions. The thick blue and red lines are steps detected from the noisy data. From these steps, we can determine the time spent in each state and, ultimately, ask basic questions about the molecule's conformational dynamics. See our J. Phys. Chem. B for details.

Trajectory probabilities predicted from path entropy maximization: Taken from our J. Chem. Phys . Much like maximum entropy in statistical mechanics predicts equilibrium state probabilities, path entropy maximization predicts distributions over dynamical paths. See our publications tab for many more papers on this topic.

Crowded nucleus: Taken from our Biophys. J. . a) Labeled BZip domain (mCherry RFP) of a transcription factor in a mouse (pituitary GHFT1) cell (Cerulean-CTA FP was used to image the cytosol). b) We can use FCS data to understand the diffusion of transcription factors in nuclear environments such as relatively free environments [A] as well as environments with crowding [B], non-specific DNA binding [C] as well as specific DNA binding [D].

Bubble boys: Preliminary fun (unpublished) data we've collected showing bacterial predators trapped in water bubbles. The predators exhibit approximately circular trajectories at the air-water interface.