app.py

import streamlit as st
from stmol import showmol
import py3Dmol
import requests
import biotite.structure.io as bsio
import random
import hashlib
import urllib3
from Bio.Blast import NCBIWWW, NCBIXML
from Bio.Seq import Seq
from Bio.SeqRecord import SeqRecord
import time
import urllib.parse

urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)

st.set_page_config(layout='wide')
st.sidebar.title('🔮 GenPro2 Protein AI Generator, Structure Predictor, and Analysis Tool')
st.sidebar.write('beta v2.12 | ')

def generate_sequence_from_words(words, length):
    seed = ' '.join(words).encode('utf-8')
    random.seed(hashlib.md5(seed).hexdigest())
    amino_acids = "ACDEFGHIKLMNPQRSTVWY"
    return ''.join(random.choice(amino_acids) for _ in range(length))

def render_mol(pdb):
    pdbview = py3Dmol.view(width=800, height=500)
    pdbview.addModel(pdb, 'pdb')
    pdbview.setStyle({'cartoon': {'color': 'spectrum'}})
    pdbview.setBackgroundColor('white')
    pdbview.zoomTo()
    pdbview.zoom(2, 800)
    pdbview.spin(True)
    showmol(pdbview, height=500, width=800)

def perform_blast_analysis(sequence):
    st.subheader('Protein Analysis')
    with st.spinner("Analyzing generated protein... This may take several minutes. Stay tuned!"):
        progress_bar = st.progress(0)
        for i in range(100):
            progress_bar.progress(i + 1)
            time.sleep(0.4)  # Analysis time
        
        try:
            record = SeqRecord(Seq(sequence), id='random_protein')
            result_handle = NCBIWWW.qblast("blastp", "swissprot", record.seq)
            
            blast_record = NCBIXML.read(result_handle)
            
            if blast_record.alignments:
                alignment = blast_record.alignments[0]  # Get the top hit
                hsp = alignment.hsps[0]  # Get the first (best) HSP
                
                # Extract protein name and organism
                title_parts = alignment.title.split('|')
                protein_name = title_parts[-1].strip()
                organism = title_parts[-2].split('OS=')[-1].split('OX=')[0].strip()
                
                # Calculate identity percentage
                identity_percentage = (hsp.identities / hsp.align_length) * 100
                
                st.write(f"**Top Match:** {protein_name}")
                st.write(f"**Organism:** {organism}")
                st.write(f"**Sequence Identity Match:** {identity_percentage:.2f}%")
                
                # Fetch protein function (if available)
                if hasattr(alignment, 'description') and alignment.description:
                    st.write(f"**Potential Function:** {alignment.description}")
                else:
                    st.write("No significant matches found in the database. This might be a unique protein sequence!")

        except Exception as e:
            st.error(f"An error occurred during protein analysis: {str(e)}")
            st.write("Please try again later, BLAST servers could be experiencing a delay.")
            
def update(sequence, word1, word2, word3, sequence_length):
    headers = {
        'Content-Type': 'application/x-www-form-urlencoded',
    }
    try:
        response = requests.post('https://api.esmatlas.com/foldSequence/v1/pdb/', 
                                 headers=headers, 
                                 data=sequence, 
                                 verify=False,
                                 timeout=300)
        response.raise_for_status()
        pdb_string = response.content.decode('utf-8')
        
        with open('predicted.pdb', 'w') as f:
            f.write(pdb_string)
        
        struct = bsio.load_structure('predicted.pdb', extra_fields=["b_factor"])
        b_value = round(struct.b_factor.mean(), 2)
        
        st.session_state.structure_info = {
            'pdb_string': pdb_string,
            'b_value': b_value,
            'word1': word1,
            'word2': word2,
            'word3': word3,
            'sequence_length': sequence_length
        }
        
        st.session_state.show_analyze_button = True

    except requests.exceptions.RequestException as e:
        st.error(f"An error occurred while calling the API: {str(e)}")
        st.write("Please try again later or contact support if the issue persists.")

def share_on_twitter(word1, word2, word3, length, plddt):
    tweet_text = f"I just generated a new protein using #GenPro2 by @WandsAI from the seed-words '#{word1}', '#{word2}', and '#{word3}' + sequence length of {length}. My Proteins plDDT Score is: {plddt}%!"
    tweet_url = f"https://twitter.com/intent/tweet?text={urllib.parse.quote(tweet_text)}"
    return tweet_url

# Initialize session state variables
if 'sequence' not in st.session_state:
    st.session_state.sequence = None
if 'show_analyze_button' not in st.session_state:
    st.session_state.show_analyze_button = False
if 'structure_info' not in st.session_state:
    st.session_state.structure_info = None

st.title("🔮 GenPro2 Protein Discovery & Anlaysis")

st.sidebar.subheader("Generate Sequence from Words")
word1 = st.sidebar.text_input("Word 1")
word2 = st.sidebar.text_input("Word 2")
word3 = st.sidebar.text_input("Word 3")
sequence_length = st.sidebar.number_input("Sequence Length", min_value=50, max_value=400, value=100, step=10)

st.markdown("""
    ## About:
GenPro2 is an end-to-end protein sequence generator, structure predictor, and analysis that uses [*ESMFold*](https://esmatlas.com/about) and the ESM-2 language model.
   

     ## How to get started:
    """)
# Information for users
st.info("""
Protein Length Guide:
- 50-100 amino acids: Small proteins/peptides
- 100-300 amino acids: Average protein domains
- 300-500 amino acids: Larger single-domain proteins

""")

st.markdown("""
1. Start by entering any three seed words of your choice and select a sequence length in the sidebar.
2. Click 'Generate and Predict' to generate a unique protein sequence based on your inputs.
3. GenPro2 then predicts the 3D structure of your protein and provides a confidence score.

GenPro2 offers a unique, hands-on way to generate, visualize, and analyze novel protein sequences. By generating and analyzing novel protein structures, you're stepping into the world of computational drug discovery. Your unique protein could be the key to unlocking new therapeutic possibilities or understanding disease mechanisms. Whether you're a student, PhD researcher, or enthusiast, GenPro2 offers a user-friendly platform to explore the vast potential of protein structures. Who knows? Your next generated sequence might just lead to a breakthrough.
    
    
    """)

if st.sidebar.button('Generate and Predict'):
    if word1 and word2 and word3:
        sequence = generate_sequence_from_words([word1, word2, word3], sequence_length)
        st.session_state.sequence = sequence
        st.sidebar.text_area("Generated Sequence", sequence, height=100)
        st.sidebar.info("Note: The same words and sequence length will always produce the same sequence.")
        
        with st.spinner("Predicting protein structure... This may take a few minutes."):
            update(sequence, word1, word2, word3, sequence_length)
    else:
        st.sidebar.warning("Please enter all three words to generate a sequence.")

# Display structure information if available
if st.session_state.structure_info:
    info = st.session_state.structure_info
    st.subheader(f'Predicted protein structure using seed: {info["word1"]}, {info["word2"]}, and {info["word3"]} + sequence length {info["sequence_length"]}')
    render_mol(info['pdb_string'])
    
    st.subheader('plDDT Confidence Score')
    st.write('plDDT is a bench mark for scoring the confidence level of protein folding predictions based on a scale from 0-100%. 70% or more is good!')
    plddt_score = int(info["b_value"] * 100)
    st.info(f'Your plDDT score is: {plddt_score}%')
 
    st.subheader("Share your unique protein on X(Twitter)")
    
    st.markdown("""
    <div style='background-color: #e6f2ff; padding: 10px; border-radius: 5px; font-size: 0.8em;'>
    <ol>
        <li>Take a screenshot of the protein structure above.</li>
        <li>Click the 'Share Results' link below to open a pre-filled post with your proteins seed-words and plDDT score.</li>
        <li>Be sure to attach a screenshot of your protein before you post!</li>
    </ol>
    </div>
    """, unsafe_allow_html=True)
    
    tweet_url = share_on_twitter(info["word1"], info["word2"], info["word3"], info["sequence_length"], plddt_score)
    st.markdown(f"[Share Results]({tweet_url})")

    st.markdown("""
    ## What to do next:

    """)

    col1, col2 = st.columns(2)
    with col1:
        if st.button('Analyze Protein'):
            perform_blast_analysis(st.session_state.sequence)
    
    with col2:
        st.download_button(
            label="Download PDB",
            data=info['pdb_string'],
            file_name='predicted.pdb',
            mime='text/plain',
        )
        
    st.markdown("""
    If you discover an interesting protein structure, you can explore it even further:
   
    1. Click the 'analyze protein' button to search the [BLAST](https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastp&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome) protein database and see if your protein matches any known sequences.  The sequence identity will show how close your sequence matches. *Note this can take several minutes
    2. Download your protein data and visit the [Protein Data Bank (PDB)](https://www.rcsb.org/) to match your protein structure against known protein structures.
    3. If you think you've discovered a new and useful protein for humans message us!

  
    **Remember, this folding is based on randomly generated sequences. Interpret the results with caution.
    Enjoy exploring the world of protein sequences!
    """)