Visiting Student-DOE CSGF– DSL – Niu, Zijian (Will)– 4.10.26

Argonne National Laboratory • Lemont, IL

6d

About The Position

Proteins are intricate molecular machines responsible for many of life's essential functions. Computational protein design now allows scientists to model and engineer novel protein sequences from physical principles or learned representations of sequence-structure-function relationships. A central goal of this field is creating binders with high affinity and specificity for a chosen target, ranging from small molecules to protein complexes. State-of-the-art approaches generally involve defining a target molecule (e.g., protein, small molecule), creating a structural scaffold around that molecule with appropriate three-dimensional complementarity, and applying an inverse folding model to generate a sequence that folds into such a structure while optimizing molecular contacts for high affinity. While recent advances have enabled de novo design of binders against proteins with nanomolar affinities, developing a general framework for small-molecule or short-peptide targets remains a major unsolved challenge. The key limitation lies in the reduced surface area of such targets, restricting the number of stabilizing interactions that can drive high-affinity binding. Moreover, current computational frameworks typically optimize binding to the desired target but lack mechanisms to penalize binding to closely related, adversarial molecules—a critical feature for achieving true biochemical specificity. To address these limitations, we propose a closed-loop framework for small-molecule and peptide binder design that couples sequence generation, molecular simulation, experimental validation, and model refinement. Using xTrimoPGLM, a pretrained protein language model (PLM), we will generate tens of thousands of candidate sequences and evaluate their binding free energies through molecular dynamics (MD) simulations with both the intended target and structurally similar adversaries. These results will be used to fine-tune the PLM via direct preference optimization (DPO), aligning its sequence-generation preferences to favor high affinity for the target and low affinity for off-targets. We will also perform experimental validation of promising candidates by measuring binding affinity via surface plasmon resonance (SPR). Iterating this loop across simulation and experimental cycles will yield a continuously improving model for high-affinity, high-specificity binder design.

Requirements

• The entirety of the appointment must be conducted within the United States.
• Must be 18 years or older at the time the appointment begins.
• Applicants must be: Currently enrolled in undergraduate or graduate studies at an accredited institution. Graduated from an accredited institution within the past 3 months; or Actively enrolled in a graduate program at an accredited institution

What This Job Offers

• Job Type: Full-time
• Career Level: Intern
• Education Level: No Education Listed
• Number of Employees: 1,001-5,000 employees