AI-Assisted Protein Analysis: From Sequence Representation to Drug-Binding Hypothesis Generation

Author: Debabrata Pruseth
Publication Date: 2026/04/26
Document Type: Technical Note / Research Article
Language: English

Abstract

Artificial intelligence and computational biology tools are making early-stage structure-based drug discovery workflows more accessible to learners, researchers, and applied AI practitioners. This paper presents an AI-assisted protein analysis pipeline that begins with protein sequence representation and ends with a cautious drug-binding hypothesis. The case study uses E. coli DHFR / folA, a 159-residue enzyme involved in folate metabolism and DNA synthesis. The author developed a workflow that uses ColabFold / AlphaFold-style structure prediction to generate a three-dimensional protein model, evaluates local structural confidence through pLDDT scores, performs exploratory molecular docking using AutoDock Vina, identifies candidate ligand-proximal residues, and uses an LLM-assisted interpretation step to convert computational outputs into a structured scientific summary. The predicted DHFR structure showed high local confidence, with a mean pLDDT of 95.49 and 146 out of 159 residues above 90. Docking produced a best AutoDock Vina score of approximately −5.686 kcal/mol, suggesting a modest but computationally plausible ligand-protein interaction. The study concludes that AI-assisted protein analysis can support early hypothesis generation, but docking scores and predicted structures require further computational refinement and experimental validation before biological claims can be made.

Keywords
AlphaFold, ColabFold, Protein Structure Prediction, DHFR, folA, AutoDock Vina, Molecular Docking, Drug Discovery, Structure-Based Drug Discovery, pLDDT, Computational Biology, Large Language Models, AI-Assisted Interpretation, Binding Hypothesis, Bioinformatics

Download Research PDF

AI-Assisted Protein Analysis: From Sequence Representation to Drug-Binding Hypothesis Generation

Suggested Citation
Pruseth, D. (2026). AI-Assisted Protein Analysis: From Sequence Representation to DrugBinding Hypothesis Generation. Debabrata Pruseth AI Blog.

Companion Note
This page provides the abstract and full-text PDF for the research version of the article. A companion blog post explains the same work in a more narrative and implementation-focused style.

Read the companion blog:
https://debabratapruseth.com/ai-assisted-protein-analysis-from-sequence-to-drug-binding-hypothesis/