Spaces:
Sleeping
Sleeping
Add requirements.txt and app.py
Browse files- app.py +144 -0
- requirements.txt +7 -0
app.py
ADDED
|
@@ -0,0 +1,144 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import joblib
|
| 3 |
+
from concurrent.futures import ThreadPoolExecutor
|
| 4 |
+
from transformers import AutoTokenizer, AutoModel, EsmModel
|
| 5 |
+
import torch
|
| 6 |
+
import numpy as np
|
| 7 |
+
import random
|
| 8 |
+
import tensorflow as tf
|
| 9 |
+
import os
|
| 10 |
+
from keras.layers import TFSMLayer
|
| 11 |
+
|
| 12 |
+
print(f"TensorFlow Version: {tf.__version__}")
|
| 13 |
+
|
| 14 |
+
base_dir = "."
|
| 15 |
+
|
| 16 |
+
# Set random seed
|
| 17 |
+
SEED = 42
|
| 18 |
+
np.random.seed(SEED)
|
| 19 |
+
random.seed(SEED)
|
| 20 |
+
torch.manual_seed(SEED)
|
| 21 |
+
if torch.cuda.is_available():
|
| 22 |
+
torch.cuda.manual_seed(SEED)
|
| 23 |
+
torch.cuda.manual_seed_all(SEED)
|
| 24 |
+
|
| 25 |
+
# Ensure deterministic behavior
|
| 26 |
+
torch.backends.cudnn.deterministic = True
|
| 27 |
+
torch.backends.cudnn.benchmark = False
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
def load_model(model_path):
|
| 31 |
+
print(f"Loading model from {model_path}...")
|
| 32 |
+
#print(f"Loading model from {model_path} using TFSMLayer...")
|
| 33 |
+
#return TFSMLayer(model_path, call_endpoint="serving_default")
|
| 34 |
+
#return tf.keras.models.load_model(model_path)
|
| 35 |
+
return tf.saved_model.load(model_path)
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
|
| 39 |
+
# Load Random Forest models and configurations
|
| 40 |
+
print("Loading models...")
|
| 41 |
+
plant_models = {
|
| 42 |
+
"Specificity": {"model": joblib.load("Specificity.pkl"), "esm_model": "facebook/esm1b_t33_650M_UR50S", "layer": 6},
|
| 43 |
+
"kcatC": {"model": joblib.load("kcatC.pkl"), "esm_model": "facebook/esm2_t36_3B_UR50D", "layer": 11},
|
| 44 |
+
"KC": {"model": joblib.load("KC.pkl"), "esm_model": "facebook/esm1b_t33_650M_UR50S", "layer": 4},
|
| 45 |
+
}
|
| 46 |
+
|
| 47 |
+
general_models = {
|
| 48 |
+
"Specificity": {"model": load_model(f"Specificity"), "esm_model": "facebook/esm2_t33_650M_UR50D", "layer": 33},
|
| 49 |
+
"kcatC": {"model": load_model(f"kcatC"), "esm_model": "facebook/esm2_t12_35M_UR50D", "layer": 7},
|
| 50 |
+
"KC": {"model": load_model(f"KC"), "esm_model": "facebook/esm2_t30_150M_UR50D", "layer": 26},
|
| 51 |
+
}
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
# Function to generate embeddings
|
| 55 |
+
def get_embedding(sequence, esm_model_name, layer):
|
| 56 |
+
print(f"Generating embeddings using {esm_model_name}, Layer {layer}...")
|
| 57 |
+
tokenizer = AutoTokenizer.from_pretrained(esm_model_name)
|
| 58 |
+
model = EsmModel.from_pretrained(esm_model_name, output_hidden_states=True)
|
| 59 |
+
|
| 60 |
+
# Tokenize the sequence
|
| 61 |
+
inputs = tokenizer(sequence, return_tensors="pt", truncation=True, max_length=1024)
|
| 62 |
+
|
| 63 |
+
# Generate embeddings
|
| 64 |
+
with torch.no_grad():
|
| 65 |
+
outputs = model(**inputs)
|
| 66 |
+
hidden_states = outputs.hidden_states # Retrieve all hidden states
|
| 67 |
+
embedding = hidden_states[layer].mean(dim=1).numpy() # Average pooling
|
| 68 |
+
|
| 69 |
+
return embedding
|
| 70 |
+
|
| 71 |
+
|
| 72 |
+
def predict_with_gpflow(model, X):
|
| 73 |
+
# Convert input to TensorFlow tensor
|
| 74 |
+
X_tensor = tf.convert_to_tensor(X, dtype=tf.float64)
|
| 75 |
+
|
| 76 |
+
# Get predictions
|
| 77 |
+
predict_fn = model.predict_f_compiled
|
| 78 |
+
mean, variance = predict_fn(X_tensor)
|
| 79 |
+
|
| 80 |
+
# Return mean and variance as numpy arrays
|
| 81 |
+
return mean.numpy().flatten(), variance.numpy().flatten()
|
| 82 |
+
# Function to predict based on user choice
|
| 83 |
+
def predict(sequence, prediction_type):
|
| 84 |
+
# Select the appropriate model set
|
| 85 |
+
selected_models = plant_models if prediction_type == "Plant-Specific" else general_models
|
| 86 |
+
|
| 87 |
+
def process_target(target):
|
| 88 |
+
esm_model_name = selected_models[target]["esm_model"]
|
| 89 |
+
layer = selected_models[target]["layer"]
|
| 90 |
+
model = selected_models[target]["model"]
|
| 91 |
+
|
| 92 |
+
# Generate embedding
|
| 93 |
+
embedding = get_embedding(sequence, esm_model_name, layer)
|
| 94 |
+
|
| 95 |
+
if prediction_type == "Plant-Specific":
|
| 96 |
+
# Random Forest prediction
|
| 97 |
+
prediction = model.predict(embedding)[0]
|
| 98 |
+
return target, round(prediction, 2)
|
| 99 |
+
else:
|
| 100 |
+
# GPflow prediction
|
| 101 |
+
mean, variance = predict_with_gpflow(model, embedding)
|
| 102 |
+
return target, round(mean[0], 2), round(variance[0], 2)
|
| 103 |
+
|
| 104 |
+
# Predict for all targets in parallel
|
| 105 |
+
with ThreadPoolExecutor() as executor:
|
| 106 |
+
results = list(executor.map(process_target, selected_models.keys()))
|
| 107 |
+
|
| 108 |
+
# Format results
|
| 109 |
+
if prediction_type == "Plant-Specific":
|
| 110 |
+
formatted_results = [
|
| 111 |
+
["Specificity", results[0][1]],
|
| 112 |
+
["kcat\u1d9c", results[1][1]],
|
| 113 |
+
["K\u1d9c", results[2][1]],
|
| 114 |
+
]
|
| 115 |
+
else:
|
| 116 |
+
formatted_results = [
|
| 117 |
+
["Specificity", results[0][1], results[0][2]],
|
| 118 |
+
["kcat\u1d9c", results[1][1], results[1][2]],
|
| 119 |
+
["K\u1d9c", results[2][1], results[2][2]],
|
| 120 |
+
]
|
| 121 |
+
|
| 122 |
+
return formatted_results
|
| 123 |
+
|
| 124 |
+
# Define Gradio interface
|
| 125 |
+
print("Creating Gradio interface...")
|
| 126 |
+
interface = gr.Interface(
|
| 127 |
+
fn=predict,
|
| 128 |
+
inputs=[
|
| 129 |
+
gr.Textbox(label="Input Protein Sequence"), # Input: Text box for sequence
|
| 130 |
+
gr.Radio(choices=["Plant-Specific", "General"], label="Prediction Type", value="Plant-Specific"), # Dropdown for selection
|
| 131 |
+
],
|
| 132 |
+
outputs=gr.Dataframe(
|
| 133 |
+
headers=["Target", "Prediction", "Uncertainty (for General)"],
|
| 134 |
+
type="array"
|
| 135 |
+
), # Output: Table
|
| 136 |
+
title="Rubisco Kinetics Prediction",
|
| 137 |
+
description=(
|
| 138 |
+
"Enter a protein sequence to predict Rubisco kinetics properties (Specificity, kcat\u1d9c, and K\u1d9c). "
|
| 139 |
+
"Choose between 'Plant-Specific' (Random Forest) or 'General' (GPflow) predictions."
|
| 140 |
+
),
|
| 141 |
+
)
|
| 142 |
+
|
| 143 |
+
if __name__ == "__main__":
|
| 144 |
+
interface.launch()
|
requirements.txt
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
transformers
|
| 2 |
+
torch
|
| 3 |
+
gradio
|
| 4 |
+
joblib
|
| 5 |
+
numpy
|
| 6 |
+
scikit-learn
|
| 7 |
+
gpflow
|