Week 5 Homework

Part A (From Pranam)

What we are doing is testing binders that neutralize the A4V mutation for SOD1. We do this by first generating peptides that could bond to SOD1. We take the A4V-mutated SOD1 sequence, add the specific binder in the format of SOD1-sequence:binder-sequence , as input for AlphaFold2-v3, then calculate the ipTM scores. These scores measure the relative confidence of the binding region.

SRWPVYAGVVEAKKX

WRWGPYAARVKLRAK

AWWGEVAGAKKWKWRAX

WRWGEYVGAVKAAAK

FLYRWPSRRGG

The experiment aimed to identify peptide binders that could neutralize the A4V mutation in SOD1 by assessing their binding confidence using AlphaFold2-v3. The pseudo-perplexity scores provided a relative measure of sequence novelty, with lower values suggesting better stability. Among the tested binders, WRWGEYVGAVKAAAK had the lowest pseudo-perplexity (10.38), suggesting a more stable binding conformation. The ipTM scores, which evaluate inter-protein structural confidence, varied across the candidates, with SRWPVYAGVVEAKK showing the highest ipTM, indicating a potentially stronger binding interface. The structural models suggest regions of high-confidence binding, with some flexibility at terminal regions.

Part B (Final Project: L-Protein Mutants)

Step 1: Information about Lysis Protein

Step 2: Mutagenesis using Protein Language Models (easier one)

Step 3: Filter and Rank your sequences with models

Step 4: 5 mutated sequences