Bioinformatics  ·  Functional Annotation  ·  University of Delaware

Playing protein detective

What does this mystery protein do? I compared its building blocks, predicted its shape, and checked published research for clues.

Bioinformatics PSI-BLAST HHpred AlphaFold2 Sequence Analysis Python
PRR1
Pseudomonas Phage
Virus that infects bacteria
gp1
Target Protein
Protein with an unclear job
4
Methods Used
Four angles on one question
✓
Possible role explored
Clues, not proof

Meet the mystery protein

Proteins are tiny biological workers. Some build things; others help reactions happen. For many proteins, we still don’t know the job.

This project looks at gp1, a protein from a bacteriophage, a virus that infects bacteria, called PRR1.

My question: what can we learn about gp1’s job by comparing it with proteins scientists already understand?

The catch: A simple match can miss distant similarities. I used several tools to look at the puzzle from different angles.

Methods

Four ways to look for clues

I looked for similar building blocks, possible shape matches, and published research. Each adds a clue; none proves the answer alone.

🔍
Step 01
Find similar sequences
Search protein databases for similar building-block sequences
PSI-BLAST
🧬
Step 02
Look for distant matches
Look for distant protein relationships that a basic search might miss
HHpred
🏗️
Step 03
Predict the shape
Predict how the protein might fold into a 3D shape
AlphaFold2
📚
Step 04
Check the research
Compare the clues with published research
Literature Review
Sequence Similarity Search

Looking for familiar patterns

PSI-BLAST searches for proteins with similar building-block sequences. In this chart, taller bars mean a closer sequence match or stronger statistical evidence, depending on the bar color.

Convergent Evidence

Putting the clues together

Clues are more useful when several approaches agree. Here’s how I compared them.

🔗
Sequence Analysis
Similar building-block patterns
Sequence searches look for familiar patterns in the protein’s building blocks. Matches can suggest a shared role, but need careful checking.
🧩
Structural Prediction
Clues from the shape
HHpred looks for distant relationships. AlphaFold2 predicts a possible 3D shape. I compared these clues with proteins whose jobs are better understood.
📖
Literature Context
Context from published research
I checked published studies to see whether the proposed role made biological sense. Predictions still need experimental testing.
Sequence Composition

What’s this protein made of?

Amino acids are a protein’s building blocks. This chart groups them by properties such as electrical charge and how they interact with water.

What I took away

My biggest lessons from a very small mystery.

01
Use more than one clue
A sequence match, a shape prediction, and a research paper each tell part of the story. Comparing them helps spot gaps.
02
Shape adds another angle
Proteins can have similar shapes even when their building-block sequences differ. Looking at both broadens the search.
03
A clue is still a clue
Software gives me matches and scores. I still need to weigh the evidence, and say clearly what remains unknown.