import py3Dmol from IPython.display import display, HTML import torch from torchvision import transforms from torch.utils.data import DataLoader, Dataset from PIL import Image import os def create_viewer_html(pdb_data, binding_residues, sequence, shift_by): binding_residues_str = ','.join([f'"{res}"' for res in binding_residues]) sequence_html = ''.join([f'

' f'{res}
' f'{i+1+shift_by if (i+1+shift_by) % 5 == 0 else " "}' f'

' for i, res in enumerate(sequence)]) html = f"""

        {sequence_html}
    

Highlight Binding Residues

""" return html def show_viewer(pdb_file_path, binding_residues, sequence, shift_by): with open(pdb_file_path, 'r') as f: pdb_data = f.read() html = create_viewer_html(pdb_data, binding_residues, sequence, shift_by) display(HTML(html)) class ImageDataset(Dataset): def __init__(self, image_dir, transform=None): self.image_dir = image_dir self.image_paths = [os.path.join(image_dir, filename) for filename in os.listdir(image_dir)] self.transform = transform def __len__(self): return len(self.image_paths) def __getitem__(self, idx): image_path = self.image_paths[idx] image = Image.open(image_path).convert('L') # Convert to grayscale if self.transform: image = self.transform(image) return image, os.path.basename(image_path) def predict_binding_residues(model, image_dir): transform = transforms.Compose([ transforms.ToTensor() ]) labels = { '0': 'nonbinding', '1': 'binding' } dataset = ImageDataset(image_dir, transform=transform) dataloader = DataLoader(dataset, batch_size=1, shuffle=False) model.eval() # Ensure model is in eval mode predictions = [] filenames = [] with torch.no_grad(): for images, paths in dataloader: outputs = model(images) _, predicted = torch.max(outputs, 1) predictions.append(predicted.item()) filenames.extend(paths) print('\rPredicting:', labels[str(predicted.item())], 'for:', paths[0], end='') binding_residues = [filename for filename, prediction in zip(filenames, predictions) if prediction == 1] return binding_residues

.css-15w88e5{color:var(--chakra-colors-fg-neutral-primary);font-weight:inherit;letter-spacing:-0.09px;}Visual-PepBDB

.css-12p60n7{color:#6B87A2;font-weight:inherit;letter-spacing:-0.09px;}Predicting & Visualizing Peptide-Protein Binding Sites with CNNs

Your PDB file:

Visual-PepBDB

Predicting & Visualizing Peptide-Protein Binding Sites with CNNs