roa7n/knots_protbertBFD_alphafold

Knots ProtBert-BFD AlphaFold

Fine-tuned ProtBert-BFD to classify proteins as knotted vs. unknotted.

Model Details

• Model type: Bert
• Language: proteins (amino acid sequences)
• Finetuned from model: Rostlab/prot_bert_bfd

Model Sources:

• Repository: CEITEC
• Paper: TBD

Usage

Dataset format:

id,sequence,label
A0A2W5F4Z7,MGGIFRVNTYYTDLEPYLQSTKLPIYGALLDGENIYELVDKSKGILVIGNESKGIRSTIQNFIQKPITIPRIGQAESLNAAVATGIIVGQLTL,1
...

Load the dataset:

import pandas as pd
from datasets import Dataset, load_dataset

df = pd.read_csv(INPUT, sep=',')
dss = Dataset.from_pandas(df)

Predict:

import torch
import numpy as np
from transformers import AutoModelForSequenceClassification, AutoTokenizer, Trainer, TrainingArguments
from math import exp

def tokenize_function(s):
    seq_split = ' '.join(s['Sequence'])
    return tokenizerM1(seq_split)

tokenizer = AutoTokenizer.from_pretrained('roa7n/knots_protbertBFD_alphafold')
model = AutoModelForSequenceClassification.from_pretrained('roa7n/knots_protbertBFD_alphafold')

tokenized_dataset = dss.map(tokenize_function, num_proc=4)
tokenized_dataset.set_format('pt')
tokenized_dataset

training_args = TrainingArguments(<PATH>, fp16=True, per_device_eval_batch_size=50, report_to='none')  

trainer = Trainer(
    model,
    training_args,
    train_dataset=tokenized_dataset,
    eval_dataset=tokenized_dataset,
    tokenizer=tokenizerM1
)

predictions, _, _ = trainer.predict(tokenized_dataset)
predictions = [np.exp(p[1]) / np.sum(np.exp(p), axis=0) for p in predictions]
df['preds'] = predictions

Evaluation

Per protein family metrics:

M1 ProtBert-BFD	Dataset size	Unknotted set size	Accuracy	TPR	TNR
All	39412	19718	0.9845	0.9865	0.9825
SPOUT	7371	550	0.9887	0.9951	0.9090
TDD	612	24	0.9901	0.9965	0.8333
DUF	716	429	0.9748	0.9721	0.9766
AdoMet synthase	1794	240	0.9899	0.9929	0.9708
Carbonic anhydrase	1531	539	0.9588	0.9737	0.9313
UCH	477	125	0.9056	0.9602	0.7520
ATCase/OTCase	3799	3352	0.9994	0.9977	0.9997
ribosomal-mitochondrial	147	41	0.8571	1.0000	0.4878
membrane	8225	1493	0.9811	0.9904	0.9390
VIT	14262	12555	0.9872	0.9420	0.9933
biosynthesis of lantibiotics	392	286	0.9642	0.9528	0.9685

Citation [optional]

BibTeX: TODO

Model Authors

Simecek: simecek@mail.muni.cz Klimentova: vae@mail.muni.cz Sramkova: denisa.sramkova@mail.muni.cz