Max Staller

Closed (1) Simulating the 3D conformations of intrinsically disordered transcriptional activation domains

Applications for fall 2021 are now closed for this project.

It is now clear that roughly a third of the amino acids in human proteins are intrinsically disordered and do not fold into a single 3D structure. Some of these regions transiently fold when bound to partners, but many simply wiggle between many confirmations. These disordered regions cannot be visualized by most structural biology methods and are difficult to simulate with traditional molecular dynamics. Instead, we use all atom Monte Carlo simulations that been optimized for disordered sequences to simulate the different conformations these proteins can inhabit. We focus on the activation domains of transcription factors, regions that bind to coactivator protein complexes. Transcription factors contain activation domains and separate DNA binding domains. DNA binding domains are conserved, structured and can be predicted from amino acid sequence. Activation domains are intrinsically disordered and poorly conserved.


Running all atom Monte Carlo simulations on the Savio cluster. Analyzing the trajectories of these simulations on the Savio cluster. Exploratory analysis of which features of the trajectories are correlated with activity.


Qualifications: Coursework in introductory biology. Computer programming in python. Experience with command line programming or working with the Savio cluster would be very helpful but is not required.

Weekly Hours: 9-11 hrs

Off-Campus Research Site: It is possible to complete this project remotely via Zoom

Related website: http://stallerlab.com

Closed (2) Probing the evolution of transcription factor activation domains

Applications for fall 2021 are now closed for this project.

Transcription factors contain DNA binding domains and separate activation domains that bind coactivator complexes. DNA binding domains are conserved, structured and can be predicted from amino acid sequence. Activation domains are intrinsically disordered (they do not fold into a single 3D structure), poorly conserved and cannot be predicted from amino acid sequence. This project will investigate activation domains from different species of yeast to understand how activation domains evolve. We have evidence that sometimes function is conserved but sequence is not conserved and we believe that these instances will help build better machine learning models for predicting activation domains from protein sequence.


Qualifications: Comparative genomics of yeast transcription factors. Starting with a Multiple Sequence Alignment of transcription factors, look for which residues are more or less conserved in known activation domains. Later, search yeast genomes and proteomes for transcription factors with BLAST or HMMER and make new Making Multiple Sequence Alignments with the hits. Specific qualifications the students should have Coursework in introductory biology. Computer programming in python.

Weekly Hours: 9-11 hrs

Off-Campus Research Site: This project can be completed remotely on Zoom.