You can deploy this model and start earning money today!

danliu1226/
PLM-interact-650M-humanV12

Name: PLM-interact-650M-humanV12
Availability: InStock

pytorch

safetensors

protein-protein interactions

paired proteins encoding

protein language model

Views

No views yet

PLM-interact model

PLM-interact： extending protein language models to predict protein-protein interactions The preprint is available at PLM-interact and the code see github link

This model is trained on human PPIs from STRING V12. For the PPI preprocessing details, see Methods in the preprint.

Model description

PLM-interact, goes beyond a single protein, jointly encoding protein pairs to learn their relationships, analogous to the next-sentence prediction task from natural language processing. This approach provides a significant improvement in performance: Trained on human-human PPIs, PLM-interact predicts mouse, fly, worm, E. coli and yeast PPIs, with 16-28% improvements in AUPR compared with state-of-the-art PPI models. Additionally, it can detect changes that disrupt or cause PPIs and be applied to virus-host PPI prediction.

An example to predict interaction probability between proteins

python

import torch
import torch.nn as nn
from transformers import AutoModel,AutoModelForMaskedLM,AutoTokenizer
import os
import torch.nn.functional as F

class PLMinteract(nn.Module):
  def __init__(self,model_name,num_labels,embedding_size): 
    super(PLMinteract,self).__init__() 
    self.esm_mask = AutoModelForMaskedLM.from_pretrained(model_name) 
    self.embedding_size=embedding_size
    self.classifier = nn.Linear(embedding_size,1) # embedding_size 
    self.num_labels=num_labels

  def forward_test(self,features):
    embedding_output = self.esm_mask.base_model(**features, return_dict=True)
    embedding=embedding_output.last_hidden_state[:,0,:] #cls token
    embedding = F.relu(embedding)
    logits = self.classifier(embedding)
    logits=logits.view(-1)
    probability = torch.sigmoid(logits)
    return  probability

# folder_huggingface_download : the download model from huggingface, such as "danliu1226/PLM-interact-650M-humanV11"
# model_name:  the ESM2 model that PLM-interact trained
# embedding_size: the embedding size of ESM2  model

folder_huggingface_download='download_huggingface_folder/'
model_name= 'facebook/esm2_t33_650M_UR50D'
embedding_size =1280

protein1 ="EGCVSNLMVCNLAYSGKLEELKESILADKSLATRTDQDSRTALHWACSAGHTEIVEFLLQLGVPVNDKDDAGWSPLHIAASAGRDEIVKALLGKGAQVNAVNQNGCTPLHYAASKNRHEIAVMLLEGGANPDAKDHYEATAMHRAAAKGNLKMIHILLYYKASTNIQDTEGNTPLHLACDEERVEEAKLLVSQGASIYIENKEEKTPLQVAKGGLGLILKRMVEG"

protein2= "MGQSQSGGHGPGGGKKDDKDKKKKYEPPVPTRVGKKKKKTKGPDAASKLPLVTPHTQCRLKLLKLERIKDYLLMEEEFIRNQEQMKPLEEKQEEERSKVDDLRGTPMSVGTLEEIIDDNHAIVSTSVGSEHYVSILSFVDKDLLEPGCSVLLNHKVHAVIGVLMDDTDPLVTVMKVEKAPQETYADIGGLDNQIQEIKESVELPLTHPEYYEEMGIKPPKGVILYGPPGTGKTLLAKAVANQTSATFLRVVGSELIQKYLGDGPKLVRELFRVAEEHAPSIVFIDEIDAIGTKRYDSNSGGEREIQRTMLELLNQLDGFDSRGDVKVIMATNRIETLDPALIRPGRIDRKIEFPLPDEKTKKRIFQIHTSRMTLADDVTLDDLIMAKDDLSGADIKAICTEAGLMALRERRMKVTNEDFKKSKENVLYKKQEGTPEGLYL"

DEVICE = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
tokenizer = AutoTokenizer.from_pretrained(model_name) 
PLMinter= PLMinteract(model_name, 1, embedding_size)
load_model = torch.load(f"{folder_huggingface_download}pytorch_model.bin")
PLMinter.load_state_dict(load_model)

texts=[protein1, protein2]
tokenized = tokenizer(*texts, padding=True, truncation='longest_first', return_tensors="pt", max_length=1603)       
tokenized = tokenized.to(DEVICE)

PLMinter.eval()
PLMinter.to(DEVICE)
with torch.no_grad():
    probability = PLMinter.forward_test(tokenized)
    print(probability.item())

Training data

Human PPIs from STRING V12

This model has been pushed to the Hub using ****:

Repo: [More Information Needed]
Docs: [More Information Needed]

danliu1226/PLM-interact-650M-humanV12

PLM-interact model

Model description

An example to predict interaction probability between proteins

Training data

danliu1226/
PLM-interact-650M-humanV12