Bioinformatics / Protein Structure

Protein Structure
Analysis: AGER

Exploring the AGER protein using UniProt, Ensembl, ClinVar, PDB and AlphaFold2/ColabFold to connect sequence information, genetic variation and predicted protein structure.

AlphaFold2 ColabFold UniProt Ensembl ClinVar PDB Python

▶ Watch demo </> View code

AGER protein structure rendered in PyMOL

Project profile

Protein AGER

Organism Homo sapiens

UniProt Q15109

Length 404 aa

Overview

AGER, also known as the receptor for advanced glycation end products, is a cell-surface pattern-recognition receptor involved in inflammatory signalling. This project combined biological databases, API-based data retrieval and protein-structure prediction to build a connected view of the protein.

Gene: AGER
Accession: Q15109
Focus: Structure
Format: Notebook

Workflow

From biological question to structure interpretation

Sequence & protein

Retrieved protein identity, function, sequence and biological annotations.

UniProt

Gene annotation

Explored transcripts, exons, gene location and genomic context.

Ensembl

Genetic variation

Investigated missense variants and available interpretations.

ClinVar

Structural data

Compared experimentally determined structures and methods.

PDB

Structure prediction

Generated ranked models and inspected prediction confidence.

ColabFold

Project video

▶

AGER walkthrough Add your video as videos/ager-demo.mp4

A short walkthrough of the databases, notebook, structure prediction and key findings.

Structure visualisation

PyMOL cartoon representation from an experimentally available AGER structure.

Sequence coverage

Multiple-sequence-alignment coverage used by the structure-prediction workflow.

Predicted models

Ranked AlphaFold2 / ColabFold outputs

Demo code

Real methods used in the notebook

Download analysis notebook Download ColabFold notebook

import requests
from io import BytesIO
import pandas as pd

protein_accession = "Q15109"

response = requests.get(
    f"https://rest.uniprot.org/uniprotkb/{protein_accession}.tsv"
)

uniprot_tsv = response.content

protein_df = pd.read_csv(
    BytesIO(uniprot_tsv),
    delimiter="\t"
)

print(protein_df)

import requests

gene_name = "AGER"

server = "https://rest.ensembl.org"

endpoint = (
    f"/xrefs/symbol/homo_sapiens/{gene_name}?"
)

response = requests.get(
    server + endpoint,
    headers={
        "Content-Type":
        "application/json"
    }
)

gene_records = response.json()

print(gene_records)

import requests

experiment_type = {}

for pdb_id in pdb_ids:
    response = requests.get(
        f"https://data.rcsb.org/rest/v1/core/entry/{pdb_id}"
    )

    if response.status_code == 200:
        record = response.json()

        experiment_type[pdb_id] = (
            record["exptl"]
        )

print(experiment_type)

Key insights

What the project demonstrates

Database integration

Connected UniProt, Ensembl, ClinVar and PDB rather than treating each source in isolation.

Structure comparison

Compared experimental evidence with computationally predicted models.

Confidence-aware interpretation

Used prediction confidence and sequence coverage to avoid overinterpreting uncertain regions.

Scientific programming

Used Python, Requests, Pandas, Biopython, PyMOL and notebook-based workflows.