The project
Bacteriocin Length and Disorder Are Associated with Lower Pooled MIC
Vivaan explored whether the structure of a bacteriocin protein can help explain how much of it is needed to stop bacterial growth. His project focused on protein length and predicted disorder, and asked whether these patterns held across different bacteriocins. The minimum inhibitory concentration (MIC) describes the lowest concentration needed to prevent visible bacterial growth.
He built a dataset from published experiments and had to reconcile studies that reported MIC values in different ways: exact values, ranges and values above a limit. Sequences were checked against NCBI and UniProt, features were measured with IUPred3, InterProScan and Clustal Omega, and an interval regression model linked MIC ranges to structural features. Extra checks for protein family, assay type and individual studies tested whether the results held up.
How the project took shape
Build a verified dataset
Collect MIC data from published studies and verify each protein sequence against NCBI and UniProt.
Measure protein features
Use IUPred3, InterProScan and Clustal Omega to quantify length, disorder and similarity between bacteriocins.
Model and stress-test
Fit an interval regression model to MIC ranges, then test whether the pattern survives adjustment for protein family, assay type and study.
Project outcomes
- Prepared a research manuscript with five figures, now being finalised.
- Built a source-verified dataset of 47 observations from 22 bacteriocins.
- Found that longer, more disordered bacteriocins tended to have lower reported MICs in the pooled analysis.
- Showed that the protein family and testing method strongly influence results, giving a clearer, more realistic direction for future research.
- Archived a reproducible analysis on Zenodo.
Vivaan is finalising his research paper. It will be added to this page once it is ready to share.
