File size: 2,877 Bytes
29d0ef7
 
 
 
 
 
 
 
9ddd36d
29d0ef7
 
 
9ddd36d
29d0ef7
9ddd36d
29d0ef7
 
 
9ddd36d
428c239
f8c84db
 
 
428c239
f8c84db
29d0ef7
 
 
ede8220
29d0ef7
ede8220
428c239
 
 
f8c84db
ede8220
ba9167b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ede8220
428c239
f8c84db
29d0ef7
 
 
 
 
 
 
 
 
 
f8c84db
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
29d0ef7
 
 
 
 
 
 
f8c84db
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
---
license: apache-2.0
base_model: nferruz/ProtGPT2
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: AmpGPT2
  results: []
---

# AmpGPT2

AmpGPT2 is a language model capable of generating de novo antimicrobial peptides (AMPs). Generated sequences are predicted to be AMPs 95.83% of the time.

## Model description

AmpGPT2 is a fine-tuned version of [nferruz/ProtGPT2](https://huggingface.co/nferruz/ProtGPT2) based on the GPT2 Transformer architecture. 

| Training Loss | Epoch | Validation Loss | Accuracy |
|:-------------:|:-----:|:---------------:|:--------:|
| 3.7948        | 50.0  | 3.9890          | 0.4213   |

To validate the results the Antimicrobial Peptide Scanner vr.2 (https://www.dveltri.com/ascan/v2/ascan.html) was used, which is a deep learning tool specifically designed for AMP recognition.

## Training and evaluation data

AmpGPT2 was trained using 32014 AMP sequences from the Compass (https://compass.mathematik.uni-marburg.de/) database. 

## How to use AmpGPT2 

The example code below contains the ideal generation settings found while testing. 
The 'num_return_sequences' parameter specifies the amount of sequences generated. When generating more than 100 sequences at the same time, I recommend doing it in batches.
The results can then be checked with the peptide scanner.
```
from transformers import pipeline
from transformers import GPT2LMHeadModel, GPT2Tokenizer

ampgpt2 = pipeline('text-generation', model="wabu/AmpGPT2")

model_amp = GPT2LMHeadModel.from_pretrained('wabu/AmpGPT2')
tokenizer_amp = GPT2Tokenizer.from_pretrained('wabu/AmpGPT2')

amp_sequences = ampgpt2( "", do_sample=True, repetition_penalty=1.2, num_return_sequences=10, eos_token_id=0 )

for i, seq in enumerate(amp_sequences):
    sequence_identifier = f"Sequence_{i + 1}"
    sequence = seq['generated_text'].replace('','').strip()

    print(f">{sequence_identifier}\n{sequence}")
```

### Training hyperparameters and results

The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50.0

\begin{table}[h!]
    \centering
    \caption{AMP Yield Comparison between AmpGPT2 and ProtGPT2}
    \begin{tabular}{lccc}
        \toprule
        Model & Total Sequences & AMP Classified & AMP Percentage (AMP\%) \\
        \midrule
        AmpGPT2 & 10000 & 9541 & 95.41\% \\
        ProtGPT2 & 10000 & 5530 & 55.3\% \\
        \bottomrule
    \end{tabular}
    \label{tab:amp_yield}
\end{table}

| Model | Amp% | Length |
|:-------:|:-----:|:-------:|
|AmpGPT2|95.86|64.08   |
|ProtGPT2| 51.85 | 222.59 |

### Framework versions

- Transformers 4.38.0.dev0
- Pytorch 2.2.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0

The model was trained on four NVIDIA A100 GPUs.