🧬 TP53 DNA to Docking — Bioinformatics Pipeline

Gene: TP53 (Tumor Protein P53) | Accession: NM_000546
Protein: p53 — "Guardian of the Genome"
Type: Tumor Suppressor Gene | Chromosome: 17p13.1

---

📋 Table of Contents

1. Project Overview
2. Gene Background
3. Step 1 — DNA Sequence Retrieval
4. Step 2 — DNA to mRNA Conversion
5. Step 3 — BLAST Analysis
6. Step 4 — AlphaFold2 Protein Structure
7. Step 5 — GNINA Ligand Docking
8. Results Summary
9. How to Run
10. References

---

Project Overview

This project performs a complete bioinformatics analysis of the TP53 tumor suppressor gene, starting from raw DNA and ending at ligand docking simulation:

DNA Sequence → mRNA Transcription → BLAST Identification
→ AlphaFold2 Structure Prediction → GNINA Ligand Docking

---

Gene Background

Property	Details
Gene Name	TP53 (Tumor Protein P53)
Gene Type	Tumor Suppressor
NCBI Gene ID	7157
Accession	NM_000546.6
Chromosome	17p13.1
Protein	p53 — 393 amino acids
Function	Cell cycle arrest, DNA repair, Apoptosis
Cancer Link	Mutated in ~50% of all human cancers

TP53 encodes the p53 protein — a transcription factor and the most critical checkpoint in the human cell cycle. When DNA damage is detected, p53 either halts division to allow repair, or triggers programmed cell death (apoptosis) if damage is too severe. Loss of TP53 function is found in nearly half of all cancers.

---

Step 1 — DNA Sequence Retrieval

Tool: Biopython (NCBI Entrez API)
Script: dna_to_mrna.py
Platform: Google Colab

The DNA sequence of TP53 was fetched directly from NCBI using the official mRNA reference accession NM_000546.

from Bio import Entrez, SeqIO
Entrez.email = "your@email.com"
handle = Entrez.efetch(db="nucleotide", id="NM_000546", rettype="gb", retmode="text")
record = SeqIO.read(handle, "genbank")

Result:

• Sequence Length: 2591 bp
• Output file: TP53_DNA.fasta

>TP53_DNA | Homo sapiens tumor protein p53 (TP53), mRNA
AGGGAGGGAGAGAATCTTCCAGGGCCAGCTCGGGCAGCAATCAGCAGG...

---

Step 2 — DNA to mRNA Conversion

Tool: Biopython Seq.transcribe()
Script: dna_to_mrna.py
Platform: Google Colab

The DNA coding sequence was transcribed to mRNA by replacing every T → U, following the central dogma of molecular biology.

mrna_seq = dna_seq.transcribe()   # T → U

	First 40 nucleotides
DNA	AGGGAGGGAGAGAATCTTCCAGGGCCAGCTCGGGCAGCA
mRNA	AGGGAGGGAGAGAAUCUUCCAGGGCCAGCUCGGGCAGCA

• mRNA Length: 2591 nucleotides
• Output file: TP53_mRNA.fasta

The protein-coding region (CDS) was also extracted and translated:

• Protein Length: 393 amino acids
• Output file: TP53_protein.fasta

---

Step 3 — BLAST Analysis

Tool: NCBI BLASTP (online + CSV export)
Database: Non-redundant protein sequences (nr)
Platform: https://blast.ncbi.nlm.nih.gov
RID: 1MNUTJ8J014

The TP53 protein sequence (393 aa) was submitted to BLASTP to confirm identity and find homologs across species.

Top 10 BLAST Hits

Rank	Accession	% Identity	E-value	Bit Score	Query Coverage
1	XP_063556659.1	99.75%	0.0	814	99.75%
2	7XZZ_K	100.00%	0.0	814	100.00%
3	8R1F_C	100.00%	0.0	814	100.00%
4	NP_000537.3	100.00%	0.0	813	100.00%
5	AYE20617.1	99.75%	0.0	813	99.75%
6	AYE20613.1	99.75%	0.0	812	99.75%
7	AYE20623.1	99.75%	0.0	812	99.75%
8	AAA61212.1	99.75%	0.0	812	99.75%
9	XP_004058559.3	99.75%	0.0	812	99.75%
10	CAA42633.1	99.75%	0.0	812	100.00%

Total hits: 100 sequences across multiple species

Key Findings

• Top hit NP_000537.3 = cellular tumor antigen p53, Homo sapiens — 100% identity, E-value 0.0 ✅
• PDB structures 7XZZ_K and 8R1F_C hit at 100% — confirms known crystallographic structures match
• All top hits have E-value = 0.0 — statistically perfect matches
• Query length confirmed at 393 amino acids — matches expected p53 protein length

---

Step 4 — AlphaFold2 Protein Structure

Tool: ColabFold (AlphaFold2)
Notebook: ColabFold on Google Colab
Job ID: TP53_structure_7c637

The TP53 protein sequence was submitted to AlphaFold2 via ColabFold to predict its 3D structure. AlphaFold2 uses deep learning trained on the entire PDB database to predict protein folding with high accuracy.

Settings Used

Parameter	Value
num_recycles	3
template_mode	pdb100
use_amber	True (structure relaxation)
Runtime	T4 GPU (Google Colab)

Output Files Generated

File	Description
TP53_structure_7c637_unrelaxed_rank_001_...pdb	Best predicted structure
TP53_structure_7c637_plddt	Per-residue confidence plot
TP53_structure_7c637_coverage	MSA coverage plot
TP53_structure_7c637_predicted_aligned_error_v1	PAE confidence matrix
TP53_structure_7c637_pae	PAE plot image

pLDDT Confidence Score

pLDDT (predicted Local Distance Difference Test) measures per-residue confidence:

Score Range	Meaning
90 – 100	Very high confidence
70 – 90	High confidence
50 – 70	Low confidence (may be disordered)
< 50	Very low (intrinsically disordered)

---

Step 5 — GNINA Ligand Docking

Tool: GNINA v1.0.3 (deep learning molecular docking)
Platform: Google Colab (Linux)
Ligand: APR-246 (Eprenetapopt) — PubChem CID 5713517

Why APR-246?

APR-246 (also called PRIMA-1MET or Eprenetapopt) is a clinically relevant small molecule that reactivates mutant p53 by covalently binding to the DNA-binding domain and restoring its wild-type conformation. It has been evaluated in Phase III clinical trials for acute myeloid leukemia (AML).

Ligand Details

Property	Value
Name	APR-246 / Eprenetapopt
PubChem CID	5713517
Formula	C7H13NO3
Molecular Weight	159.18 g/mol
Target	TP53 DNA-binding domain
Clinical Stage	Phase III trials

Docking Configuration

./gnina \
  --receptor protein.pdb \
  --ligand APR246.sdf \
  --out docked_poses.sdf \
  --center_x 1.5 --center_y 2.3 --center_z -0.8 \
  --size_x 25 --size_y 25 --size_z 25 \
  --num_modes 9 \
  --exhaustiveness 8 \
  --cnn_scoring none

Docking Results

Mode	Affinity (kcal/mol)	CNN Pose Score	CNN Affinity
1 (Best)	-5.22	0.1708	4.485
2	-5.14	0.1462	4.487
3	-4.69	0.1393	4.901
4	-4.13	0.1357	4.409
5	-5.19	0.1119	4.308
6	-5.33	0.1021	4.233
7	-4.69	0.1007	4.598
8	-5.16	0.0949	4.726
9	-5.56	0.0862	4.626

Strongest binding affinity: -5.56 kcal/mol (Mode 9)
Mode 1 selected as best overall pose (highest CNN pose score: 0.1708)

Output file: docked_poses.sdf — contains all 9 predicted binding poses with CNN scores and binding affinities (kcal/mol).

Docking Results

Model 1

Model 2

Model 3

Model 4

Model 5

Model 6

Model 7

Model 8

Model 9

Results Summary

Step	Tool	Output	Key Result
1 — DNA retrieval	Biopython / NCBI	TP53_DNA.fasta	2591 bp sequence
2 — mRNA conversion	Biopython	TP53_mRNA.fasta	T→U transcription, 393 aa protein
3 — BLAST	NCBI BLASTP	HitTable CSV	100% identity to p53, E-value 0.0
4 — AlphaFold2	ColabFold	.pdb structure	High-confidence 3D structure predicted
5 — Docking	GNINA	docked_poses.sdf	APR-246 docked to DNA-binding domain

---

How to Run

Requirements

pip install biopython

Step 1 & 2 — DNA + mRNA

python dna_to_mrna.py

Step 3 — BLAST

Visit https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastp and paste sequence from TP53_protein.fasta

Step 4 — AlphaFold2

Open ColabFold notebook → paste protein sequence → Runtime → Run All: https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb

Step 5 — GNINA Docking (Google Colab)

# Install GNINA
wget https://github.com/gnina/gnina/releases/download/v1.0.3/gnina -O gnina
chmod +x gnina

# Run docking
./gnina --receptor protein.pdb --ligand APR246.sdf --out docked_poses.sdf \
        --center_x 1.5 --center_y 2.3 --center_z -0.8 \
        --size_x 25 --size_y 25 --size_z 25 \
        --num_modes 9 --exhaustiveness 8 --cnn_scoring none

---

References

1. NCBI Gene TP53 — https://www.ncbi.nlm.nih.gov/gene/7157
2. UniProt P04637 (p53 Human) — https://www.uniprot.org/uniprot/P04637
3. ColabFold / AlphaFold2 — https://github.com/sokrypton/ColabFold
4. GNINA Molecular Docking — https://github.com/gnina/gnina
5. APR-246 (PubChem CID 5713517) — https://pubchem.ncbi.nlm.nih.gov/compound/5713517
6. PDB Structure 2OCJ — https://www.rcsb.org/structure/2OCJ