fix app

NeuroPredPLM/predict.py

@@ -15,4 +15,17 @@ def predict(peptide_list, model_path, device='cpu'):
         att = att.cpu().numpy()
         out = {'Neuropeptide':pred[0][1], "Non-neuropeptide":pred[0][0]}
     return out
-    
+
+def batch_predict(peptide_list, cutoff, model_path, device='cpu'):
+    with torch.no_grad():
+        neuroPred_model = EsmModel()
+        neuroPred_model.eval()
+        # state_dict = load_hub_workaround(MODEL_URL)
+        state_dict = torch.load(model_path, map_location="cpu")
+        neuroPred_model.load_state_dict(state_dict)
+        neuroPred_model = neuroPred_model.to(device)
+        prob, att = neuroPred_model(peptide_list, device)
+        pred = torch.softmax(prob, dim=-1).cpu().tolist()
+        att = att.cpu().numpy()
+        out = [[i[0], i[1], f"{j[1]:.3f}", 'Neuropeptide' if j[1] >cutoff else 'Non-neuropeptide'] for i, j in zip(peptide_list, pred)]
+    return out

app.py

@@ -1,5 +1,5 @@
 import torch
-from NeuroPredPLM.predict import predict
+from NeuroPredPLM.predict import predict, batch_predict
 import gradio as gr
 from io import StringIO
 from Bio import SeqIO
@@ -14,9 +14,64 @@ def classifier(peptide_seq):
     return neuropeptide_pred
 # {peptide_id:[Type:int(1->neuropeptide,0->non-neuropeptide), attention score:nd.array]}
 
-iface = gr.Interface(fn=classifier, inputs=gr.Textbox(
-            label="Input peptide sequence",
-            lines=3,
-            value=">peptide-1\nIGLRLPNMLKF",
-        ), outputs=gr.outputs.Label(num_top_classes=2), title="NeuroPred-PLM")
-iface.launch()
+
+def batch_classifier(file, cutoff):
+    data = []
+    for record in SeqIO.parse(file.name, 'fasta'):
+        data.append((record.id, str(record.seq)))
+    device = "cuda" if torch.cuda.is_available() else "cpu" 
+    neuropeptide_pred = batch_predict(data, cutoff, './model.pth', device)
+    return neuropeptide_pred
+
+with gr.Blocks() as demo:
+    gr.Markdown(" ## NeuroPred-PLM")
+    gr.Markdown("In this work, we developed an interpretable and robust neuropeptide prediction model, named NeuroPred-PLM. First, we employed a language model (ESM) of proteins to obtain semantic representations of neuropeptides, which could reduce the complexity of feature engineering. Next, we adopted a multi-scale convolutional neural network to enhance the local feature representation of neuropeptide embeddings. To make the model interpretable, we proposed a global multi-head attention network that could be used to capture the position-wise contribution to neuropeptide prediction via the attention scores. In addition, NeuroPred-PLM was developed based on our newly constructed NeuroPep 2.0 database. Benchmarks based on the independent test set show that NeuroPred-PLM achieves superior predictive performance compared to other state-of-the-art predictors.")
+    with gr.Tab("Single Sequence Medel"):
+        # cutoff = gr.Slider(0, 1, step=0.1, value=0.5, interactive=True)
+        with gr.Row():
+            with gr.Column(scale=2):
+                text_input = gr.Textbox(
+                    label="Input single peptide sequence in the Fasta format",
+                    lines=4,
+                    value=">peptide-1\nIGLRLPNMLKF",
+                )
+                gr.Markdown("#### The input peptide sequence length should be between 5-100")
+                single_cutoff = gr.Slider(0, 1, step=0.1, value=0.5, interactive=True, label="Threshold")
+                text_button = gr.Button("Submit")
+            with gr.Column(scale=2):
+                text_output = gr.outputs.Label(num_top_classes=2, label='Output')
+    with gr.Tab("Batch Model"):
+        with gr.Row():
+            with gr.Column(scale=2):
+                input_file_fasta = gr.File()
+                # cutoff = gr.Slider(0, 1, step=0.1, value=0.5, interactive=True, label="threshold")
+                # image_button = gr.Button("Submit")
+            with gr.Column(scale=2):
+                batch_cutoff = gr.Slider(0, 1, step=0.1, value=0.5, interactive=True, label="Threshold")
+                gr.Markdown("### Note")
+                gr.Markdown("- Limit the number of input sequences to less than 100")
+                gr.Markdown("- The file should be the Fasta format")
+                gr.Markdown("- The input peptide sequence length should be between 5-100")
+                image_button = gr.Button("Submit")
+        with gr.Column():
+            # gr.Markdown(" ### Flip text or image files using this demo.")
+            frame_output = gr.DataFrame(headers=["Sequence Id", "Sequence", "Probability of neuropeptides", "Neuropeptide"],
+            datatype=["str", "str", "str", 'str'],)
+
+        
+
+    with gr.Accordion("Citation"):
+        gr.Markdown("- Wang, L., Huang, C., Wang, M., Xue, Z., & Wang, Y. (2022). NeuroPred-PLM: an interpretable and robust model for neuropeptide prediction by protein language model. In preparation.")
+        gr.Markdown("- GitHub: https://github.com/ISYSLAB-HUST/NeuroPred-PLM")
+
+    with gr.Accordion("License"):
+        gr.Markdown("- Released under the [MIT license](https://github.com/ISYSLAB-HUST/NeuroPred-PLM/blob/main/LICENSE). ")
+
+    with gr.Accordion("Contact"):
+        gr.Markdown("- If you have any questions, comments, or would like to report a bug, please file a Github issue or contact me at wanglei94@hust.edu.cn.")
+
+    text_button.click(classifier, inputs=text_input, outputs=text_output)
+    image_button.click(batch_classifier, inputs=[input_file_fasta, batch_cutoff], outputs=frame_output)
+
+demo.queue(4)
+demo.launch()

test.fa

@@ -0,0 +1,6 @@
+>peptide_1
+IGLRLPNMLKF
+>peptide_2
+QAAQFKVWSASELVD
+>peptide_3
+LRSPKMMHKSGCFGRRLDRIGSLSGLGCNVLRKY