<br/>
<h1 align="center">Ankh ☥: Optimized Protein Language Model Unlocks General-Purpose Modelling </h1>
<br/>
<br/>
[Ankh](https://arxiv.org/abs/2301.06568) is the first general-purpose protein language model trained on Google's **TPU-V4** surpassing the state-of-the-art performance with dramatically less parameters, promoting accessibility to research innovation via attainable resources.
<div align="center"><img width=500 height=350 src="https://github.com/agemagician/Ankh/blob/main/images/AnkhGIF.gif?raw=true"></div>
This repository will be updated regulary with **new pre-trained models for proteins** in part of supporting the **biotech** community in revolutinizing protein engineering using AI.
Table of Contents
=================
* [ Installation](#install)
* [ Models Availability](#models)
* [ Dataset Availability](#datasets)
* [ Usage](#usage)
* [ Original downstream Predictions](#results)
* [ Followup use-cases](#inaction)
* [ Comparisons to other tools](#comparison)
* [ Community and Contributions](#community)
* [ Have a question?](#question)
* [ Found a bug?](#bug)
* [ Requirements](#requirements)
* [ Sponsors](#sponsors)
* [ Team](#team)
* [ License](#license)
* [ Citation](#citation)
<a name="install"></a>
## Installation
```python
python -m pip install ankh
```
<a name="models"></a>
## Models Availability
| Model | ankh | Hugging Face |
|------------------------------------|-----------------------------------|-----------------------------------------------------------|
| Ankh Large | `ankh.load_large_model()` |[Ankh Large](https://huggingface.co/ElnaggarLab/ankh-large)|
| Ankh Base | `ankh.load_base_model()` |[Ankh Base](https://huggingface.co/ElnaggarLab/ankh-base) |
<a name="datasets"></a>
## Datasets Availability
| Dataset | Hugging Face |
| ----------------------------- |---------------------------------------------------------------------------------------------------|
| Remote Homology | `load_dataset("proteinea/remote_homology")` |
| CASP12 | `load_dataset("proteinea/secondary_structure_prediction", data_files={'test': ['CASP12.csv']})`|
| CASP14 | `load_dataset("proteinea/secondary_structure_prediction", data_files={'test': ['CASP14.csv']})`|
| CB513 | `load_dataset("proteinea/secondary_structure_prediction", data_files={'test': ['CB513.csv']})` |
| TS115 | `load_dataset("proteinea/secondary_structure_prediction", data_files={'test': ['TS115.csv']})` |
| DeepLoc | `load_dataset("proteinea/deeploc")` |
| Fluorescence | `load_dataset("proteinea/fluorescence")` |
| Solubility | `load_dataset("proteinea/solubility")` |
| Nearest Neighbor Search | `load_dataset("proteinea/nearest_neighbor_search")` |
<a name="usage"></a>
## Usage
* Loading pre-trained models:
```python
import ankh
# To load large model:
model, tokenizer = ankh.load_large_model()
model.eval()
# To load base model.
model, tokenizer = ankh.load_base_model()
model.eval()
```
* Feature extraction using ankh large example:
```python
model, tokenizer = ankh.load_large_model()
model.eval()
protein_sequences = ['MKALCLLLLPVLGLLVSSKTLCSMEEAINERIQEVAGSLIFRAISSIGLECQSVTSRGDLATCPRGFAVTGCTCGSACGSWDVRAETTCHCQCAGMDWTGARCCRVQPLEHHHHHH',
'GSHMSLFDFFKNKGSAATATDRLKLILAKERTLNLPYMEEMRKEIIAVIQKYTKSSDIHFKTLDSNQSVETIEVEIILPR']
protein_sequences = [list(seq) for seq in protein_sequences]
outputs = tokenizer.batch_encode_plus(protein_sequences,
add_special_tokens=True,
padding=True,
is_split_into_words=True,
return_tensors="pt")
with torch.no_grad():
embeddings = model(input_ids=outputs['input_ids'], attention_mask=outputs['attention_mask'])
```
* Loading downstream models example:
```python
# To use downstream model for binary classification:
binary_classification_model = ankh.ConvBertForBinaryClassification(input_dim=768,
nhead=4,
hidden_dim=384,
num_hidden_layers=1,
num_layers=1,
kernel_size=7,
dropout=0.2,
pooling='max')
# To use downstream model for multiclass classification:
multiclass_classification_model = ankh.ConvBertForMultiClassClassification(num_tokens=2,
input_dim=768,
nhead=4,
hidden_dim=384,
num_hidden_layers=1,
num_layers=1,
kernel_size=7,
dropout=0.2)
# To use downstream model for regression:
# training_labels_mean is optional parameter and it's used to fill the output layer's bias with it,
# it's useful for faster convergence.
regression_model = ankh.ConvBertForRegression(input_dim=768,
nhead=4,
hidden_dim=384,
num_hidden_layers=1,
num_layers=1,
kernel_size=7,
dropout=0,
pooling='max',
training_labels_mean=0.38145)
```
<a name="results"></a>
## Original downstream Predictions
<a name="q3"></a>
* <b> Secondary Structure Prediction (Q3):</b><br/>
| Model | CASP12 | CASP14 (HARD) | TS115 | CB513 |
|--------------------------|:----------------:|:-------------:|:-------------:|:------------:|
|**Ankh Large** | 83.59% | 77.48% | 88.22% | 88.48% |
|Ankh Base | 80.81% | 76.67% | 86.92% | 86.94% |
|ProtT5-XL-UniRef50 | 83.34% | 75.09% | 86.82% | 86.64% |
|ESM2-15B | 83.16% | 76.56% | 87.50% | 87.35% |
|ESM2-3B | 83.14% | 76.75% | 87.50% | 87.44% |
|ESM2-650M | 82.43% | 76.97% | 87.22% | 87.18% |
|ESM-1b | 79.45% | 75.39% | 85.02% | 84.31% |
<a name="q8"></a>
* <b> Secondary Structure Prediction (Q8):</b><br/>
| Model | CASP12 | CASP14 (HARD) | TS115 | CB513 |
|--------------------------|:----------------:|:-------------:|:-------------:|:------------:|
|**Ankh Large** | 71.69% | 63.17% | 79.10% | 78.45% |
|Ankh Base | 68.85% | 62.33% | 77.08% | 75.83% |
|ProtT5-XL-UniRef50 | 70.47% | 59.71% | 76.91% | 74.81% |
|ESM2-15B | 71.17% | 61.81% | 77.67% | 75.88% |
|ESM2-3B | 71.69% | 61.52% | 77.62% | 75.95% |
|ESM2-650M | 70.50% | 62.10% | 77.68% | 75.89% |
|ESM-1b | 66.02% | 60.34% | 73.82% | 71.55% |
<a name="CP"></a>
* <b> Contact Prediction Long Precision Using Embeddings:</b><br/>
| Model | ProteinNet (L/1) | ProteinNet (L/5) | CASP14 (L/1) | CASP14 (L/5) |
|--------------------------|:----------------:|:----------------:|:-------------:|:------------:|
|**Ankh Large** | 48.93% | 73.49% | 16.01% | 29.91% |
|Ankh Base | 43.21% | 66.63% | 13.50% | 28.65% |
|ProtT5-XL-UniRef50 | 44.74% | 68.95% | 11.95% | 24.45% |
|ESM2-15B | 31.62% | 52.97% | 14.44% | 26.61% |
|ESM2-3B | 30.24% | 51.34% | 12.20% | 21.91% |
|ESM2-650M | 29.36% | 50.74% | 13.71% | 22.25% |
|ESM-1b | 29.25% | 50.69% | 10.18% | 18.08% |
<a name="CP"></a>
* <b> Contact Prediction Long Precision Using attention scores:</b><br/>
| Model | ProteinNet (L/1) | ProteinNet (L/5) | CASP14 (L/1) | CASP14 (L/5) |
|--------------------------|:----------------:|:----------------:|:-------------:|:------------:|
|**Ankh Large** | 31.44% | 55.58% | 11.05% | 20.74% |
|Ankh Base | 25.93% | 46.28% | 9.32% | 19.51% |
|ProtT5-XL-UniRef50 | 30.85% | 51.90% | 8.60% | 16.09% |
|ESM2-15B | 33.32% | 57.44% | 12.25% | 24.60% |
|ESM2-3B | 33.92% | 56.63% | 12.17% | 21.36% |
|ESM2-650M | 31.87% | 54.63% | 10.66% | 21.01% |
|ESM-1b | 25.30% | 42.03% | 7.77% | 15.77% |
<a name="Loc"></a>
* <b> Localization (Q10):</b><br/>
| Model | DeepLoc Dataset |
|--------------------------|:----------------:|
|**Ankh Large** | 83.01% |
|Ankh Base | 81.38% |
|ProtT5-XL-UniRef50 | 82.95% |
|ESM2-15B | 81.22% |
|ESM2-3B | 81.22% |
|ESM2-650M | 82.08% |
|ESM-1b | 80.51% |
<a name="RH"></a>
* <b> Remote Homology:</b><br/>
| Model | SCOPe (Fold) |
|--------------------------|:----------------:|
|Ankh Large | 61.01% |
|**Ankh Base** | 61.14% |
|ProtT5-XL-UniRef50 | 59.38% |
|ESM2-15B | 54.48% |
|ESM2-3B | 59.24% |
|ESM2-650M | 51.36% |
|ESM-1b | 56.93% |
<a name="Sol"></a>
* <b> Solubility:</b><br/>
| Model | Solubility |
|--------------------------|:----------------:|
|**Ankh Large** | 76.41% |
|Ankh Base | 76.36% |
|ProtT5-XL-UniRef50 | 76.26% |
|ESM2-15B | 60.52% |
|ESM2-3B | 74.91% |
|ESM2-650M | 74.56% |
|ESM-1b | 74.91% |
<a name="Flu"></a>
* <b> Fluorescence (Spearman Correlation):</b><br/>
| Model | Fluorescence |
|--------------------------|:----------------:|
|**Ankh Large** | 0.62 |
|Ankh Base | 0.62 |
|ProtT5-XL-UniRef50 | 0.61 |
|ESM2-15B | 0.56 |
|ESM-1b | 0.48 |
|ESM2-650M | 0.48 |
|ESM2-3B | 0.46 |
<a name="CATH"></a>
* <b> Nearest Neighbor Search using Global Pooling:</b><br/>
| Model | Lookup69K (C) | Lookup69K (A) | Lookup69K (T) | Lookup69K (H) |
|--------------------------|:----------------:|:----------------:|:----------------:|:----------------:|
|Ankh Large | 0.83 | 0.72 | 0.60 | 0.70 |
|**Ankh Base** | 0.85 | 0.77 | 0.63 | 0.72 |
|ProtT5-XL-UniRef50 | 0.83 | 0.69 | 0.57 | 0.73 |
|ESM2-15B | 0.78 | 0.63 | 0.52 | 0.67 |
|ESM2-3B | 0.79 | 0.65 | 0.53 | 0.64 |
|ESM2-650M | 0.72 | 0.56 | 0.40 | 0.53 |
|ESM-1b | 0.78 | 0.65 | 0.51 | 0.63 |
<a name="team"></a>
## Team
* <b>Technical University of Munich:</b><br/>
| [Ahmed Elnaggar](https://github.com/agemagician) | Burkhard Rost |
|:------------------------------------------------:|:-------------------------:|
| <img width=120 src="https://github.com/agemagician/Ankh/blob/main/images/AhmedElNaggar.jpg?raw=true"> | <img width=120 src="https://github.com/agemagician/Ankh/blob/main/images/Rost.jpg?raw=true"> |
* <b>Proteinea:</b><br/>
| [Hazem Essam](https://github.com/hazemessamm) | [Wafaa Ashraf](https://github.com/wafaaashraf) | [Walid Moustafa](https://github.com/wmustafaawad) | [Mohamed Elkerdawy](https://github.com/melkerdawy) |
|:-------------------------:|:-------------------------:|:-------------------------:|:-------------------------:|
| <img width=120 src="https://github.com/agemagician/Ankh/blob/main/images/HazemEssam.jpeg?raw=true"> | <img width=120 src="https://github.com/agemagician/Ankh/blob/main/images/WafaaAshraf.jpeg?raw=true"> | <img width=120 src="https://github.com/agemagician/Ankh/blob/main/images/WalidMoustafa.jpg?raw=true"> | <img width=120 src="https://github.com/agemagician/Ankh/blob/main/images/MohamedElKerdawy.jpeg?raw=true"> |
* <b>University of Columbia:</b><br/>
| [Charlotte Rochereau](https://github.com/crochereau) |
|:----------------------------------------------------:|
| <img width=120 src="https://github.com/agemagician/Ankh/blob/main/images/CharlotteRochereau.jpg?raw=true"> |
<a name="sponsors"></a>
## Sponsors
| Google Cloud |
:------------------------------------------------------------------------------------------------------------------------:|
<img width=120 src="https://github.com/agemagician/Ankh/blob/main/images/google_cloud_logo.jpg?raw=true"> |
<a name="license"></a>
## License
Ankh pretrained models are released under the under terms of the [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International](https://creativecommons.org/licenses/by/4.0/).
<a name="community"></a>
## Community and Contributions
The Ankh project is a **open source project** supported by various partner companies and research institutions. We are committed to **share all our pre-trained models and knowledge**. We are more than happy if you could help us on sharing new ptrained models, fixing bugs, proposing new feature, improving our documentation, spreading the word, or support our project.
<a name="question"></a>
## Have a question?
We are happy to hear your question in our issues page [Ankh](https://github.com/agemagician/Ankh/issues)! Obviously if you have a private question or want to cooperate with us, you can always **reach out to us directly** via [Hello](mailto:hello@proteinea.com?subject=[GitHub]Ankh).
<a name="bug"></a>
## Found a bug?
Feel free to **file a new issue** with a respective title and description on the the [Ankh](https://github.com/agemagician/Ankh/issues) repository. If you already found a solution to your problem, **we would love to review your pull request**!.
<a name="citation"></a>
## ✏️ Citation
If you use this code or our pretrained models for your publication, please cite the original paper:
```
@article{elnaggar2023ankh,
title={Ankh: Optimized Protein Language Model Unlocks General-Purpose Modelling},
author={Elnaggar, Ahmed and Essam, Hazem and Salah-Eldin, Wafaa and Moustafa, Walid and Elkerdawy, Mohamed and Rochereau, Charlotte and Rost, Burkhard},
journal={arXiv preprint arXiv:2301.06568},
year={2023}
}
```
Raw data
{
"_id": null,
"home_page": "https://github.com/agemagician/Ankh",
"name": "ankh",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.7,<4.0",
"maintainer_email": "",
"keywords": "ankh",
"author": "Ahmed Elnaggar",
"author_email": "amit.najjar@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/4e/c1/76767b3422b9d2042403dc5e1c441695e9279e804350e333a2c5c1e7b464/ankh-1.10.0.tar.gz",
"platform": null,
"description": "<br/>\n\n<h1 align=\"center\">Ankh \u2625: Optimized Protein Language Model Unlocks General-Purpose Modelling </h1>\n<br/>\n\n<br/>\n\n[Ankh](https://arxiv.org/abs/2301.06568) is the first general-purpose protein language model trained on Google's **TPU-V4** surpassing the state-of-the-art performance with dramatically less parameters, promoting accessibility to research innovation via attainable resources.\n\n\n<div align=\"center\"><img width=500 height=350 src=\"https://github.com/agemagician/Ankh/blob/main/images/AnkhGIF.gif?raw=true\"></div>\n\n\n\n\nThis repository will be updated regulary with **new pre-trained models for proteins** in part of supporting the **biotech** community in revolutinizing protein engineering using AI.\n\n\nTable of Contents\n=================\n* [ Installation](#install)\n* [ Models Availability](#models)\n* [ Dataset Availability](#datasets)\n* [ Usage](#usage)\n* [ Original downstream Predictions](#results)\n* [ Followup use-cases](#inaction)\n* [ Comparisons to other tools](#comparison)\n* [ Community and Contributions](#community)\n* [ Have a question?](#question)\n* [ Found a bug?](#bug)\n* [ Requirements](#requirements)\n* [ Sponsors](#sponsors)\n* [ Team](#team)\n* [ License](#license)\n* [ Citation](#citation)\n\n\n<a name=\"install\"></a>\n## Installation\n\n```python\npython -m pip install ankh\n```\n\n\n<a name=\"models\"></a>\n## Models Availability\n\n| Model | ankh | Hugging Face |\n|------------------------------------|-----------------------------------|-----------------------------------------------------------|\n| Ankh Large | `ankh.load_large_model()` |[Ankh Large](https://huggingface.co/ElnaggarLab/ankh-large)| \n| Ankh Base | `ankh.load_base_model()` |[Ankh Base](https://huggingface.co/ElnaggarLab/ankh-base) |\n\n\n<a name=\"datasets\"></a>\n\n## Datasets Availability\n| Dataset | Hugging Face | \n| ----------------------------- |---------------------------------------------------------------------------------------------------|\n|\tRemote Homology \t | `load_dataset(\"proteinea/remote_homology\")` |\n|\tCASP12\t\t\t | `load_dataset(\"proteinea/secondary_structure_prediction\", data_files={'test': ['CASP12.csv']})`|\n|\tCASP14\t\t\t | `load_dataset(\"proteinea/secondary_structure_prediction\", data_files={'test': ['CASP14.csv']})`|\n|\tCB513\t\t\t | `load_dataset(\"proteinea/secondary_structure_prediction\", data_files={'test': ['CB513.csv']})` |\n|\tTS115\t\t\t | `load_dataset(\"proteinea/secondary_structure_prediction\", data_files={'test': ['TS115.csv']})` |\n|\tDeepLoc\t\t | `load_dataset(\"proteinea/deeploc\")` |\n| Fluorescence | `load_dataset(\"proteinea/fluorescence\")` |\n| Solubility | `load_dataset(\"proteinea/solubility\")` |\n| Nearest Neighbor Search | `load_dataset(\"proteinea/nearest_neighbor_search\")` |\n\n\n\n<a name=\"usage\"></a>\n## Usage\n\n* Loading pre-trained models:\n```python\n import ankh\n\n # To load large model:\n model, tokenizer = ankh.load_large_model()\n model.eval()\n\n\n # To load base model.\n model, tokenizer = ankh.load_base_model()\n model.eval()\n```\n\n* Feature extraction using ankh large example:\n```python\n\n model, tokenizer = ankh.load_large_model()\n model.eval()\n\n protein_sequences = ['MKALCLLLLPVLGLLVSSKTLCSMEEAINERIQEVAGSLIFRAISSIGLECQSVTSRGDLATCPRGFAVTGCTCGSACGSWDVRAETTCHCQCAGMDWTGARCCRVQPLEHHHHHH', \n 'GSHMSLFDFFKNKGSAATATDRLKLILAKERTLNLPYMEEMRKEIIAVIQKYTKSSDIHFKTLDSNQSVETIEVEIILPR']\n\n protein_sequences = [list(seq) for seq in protein_sequences]\n\n\n outputs = tokenizer.batch_encode_plus(protein_sequences, \n add_special_tokens=True, \n padding=True, \n is_split_into_words=True, \n return_tensors=\"pt\")\n with torch.no_grad():\n embeddings = model(input_ids=outputs['input_ids'], attention_mask=outputs['attention_mask'])\n```\n\n* Loading downstream models example:\n```python\n # To use downstream model for binary classification:\n binary_classification_model = ankh.ConvBertForBinaryClassification(input_dim=768, \n nhead=4, \n hidden_dim=384, \n num_hidden_layers=1, \n num_layers=1, \n kernel_size=7, \n dropout=0.2, \n pooling='max')\n\n # To use downstream model for multiclass classification:\n multiclass_classification_model = ankh.ConvBertForMultiClassClassification(num_tokens=2, \n input_dim=768, \n nhead=4, \n hidden_dim=384, \n num_hidden_layers=1, \n num_layers=1, \n kernel_size=7, \n dropout=0.2)\n\n # To use downstream model for regression:\n # training_labels_mean is optional parameter and it's used to fill the output layer's bias with it, \n # it's useful for faster convergence.\n regression_model = ankh.ConvBertForRegression(input_dim=768, \n nhead=4, \n hidden_dim=384, \n num_hidden_layers=1, \n num_layers=1, \n kernel_size=7, \n dropout=0, \n pooling='max', \n training_labels_mean=0.38145)\n```\n\n\n<a name=\"results\"></a>\n## Original downstream Predictions \n\n<a name=\"q3\"></a>\n * <b> Secondary Structure Prediction (Q3):</b><br/>\n \n| Model | CASP12 | CASP14 (HARD) | TS115 | CB513 |\n|--------------------------|:----------------:|:-------------:|:-------------:|:------------:|\n|**Ankh Large** | 83.59% | 77.48% | 88.22% | 88.48% |\n|Ankh Base | 80.81% | 76.67% | 86.92% | 86.94% |\n|ProtT5-XL-UniRef50 | 83.34% | 75.09% | 86.82% | 86.64% |\n|ESM2-15B | 83.16% | 76.56% | 87.50% | 87.35% |\n|ESM2-3B | 83.14% | 76.75% | 87.50% | 87.44% |\n|ESM2-650M | 82.43% | 76.97% | 87.22% | 87.18% |\n|ESM-1b | 79.45% | 75.39% | 85.02% | 84.31% |\n\n\n<a name=\"q8\"></a>\n * <b> Secondary Structure Prediction (Q8):</b><br/>\n \n| Model | CASP12 | CASP14 (HARD) | TS115 | CB513 |\n|--------------------------|:----------------:|:-------------:|:-------------:|:------------:|\n|**Ankh Large** | 71.69% | 63.17% | 79.10% | 78.45% |\n|Ankh Base | 68.85% | 62.33% | 77.08% | 75.83% |\n|ProtT5-XL-UniRef50 | 70.47% | 59.71% | 76.91% | 74.81% |\n|ESM2-15B | 71.17% | 61.81% | 77.67% | 75.88% |\n|ESM2-3B | 71.69% | 61.52% | 77.62% | 75.95% |\n|ESM2-650M | 70.50% | 62.10% | 77.68% | 75.89% |\n|ESM-1b | 66.02% | 60.34% | 73.82% | 71.55% |\n\n<a name=\"CP\"></a>\n * <b> Contact Prediction Long Precision Using Embeddings:</b><br/>\n \n| Model | ProteinNet (L/1) | ProteinNet (L/5) | CASP14 (L/1) | CASP14 (L/5) |\n|--------------------------|:----------------:|:----------------:|:-------------:|:------------:|\n|**Ankh Large** | 48.93% | 73.49% | 16.01% | 29.91% |\n|Ankh Base | 43.21% | 66.63% | 13.50% | 28.65% |\n|ProtT5-XL-UniRef50 | 44.74% | 68.95% | 11.95% | 24.45% |\n|ESM2-15B | 31.62% | 52.97% | 14.44% | 26.61% |\n|ESM2-3B | 30.24% | 51.34% | 12.20% | 21.91% |\n|ESM2-650M | 29.36% | 50.74% | 13.71% | 22.25% |\n|ESM-1b | 29.25% | 50.69% | 10.18% | 18.08% |\n\n\n<a name=\"CP\"></a>\n * <b> Contact Prediction Long Precision Using attention scores:</b><br/>\n \n| Model | ProteinNet (L/1) | ProteinNet (L/5) | CASP14 (L/1) | CASP14 (L/5) |\n|--------------------------|:----------------:|:----------------:|:-------------:|:------------:|\n|**Ankh Large** | 31.44% | 55.58% | 11.05% | 20.74% |\n|Ankh Base | 25.93% | 46.28% | 9.32% | 19.51% |\n|ProtT5-XL-UniRef50 | 30.85% | 51.90% | 8.60% | 16.09% |\n|ESM2-15B | 33.32% | 57.44% | 12.25% | 24.60% |\n|ESM2-3B | 33.92% | 56.63% | 12.17% | 21.36% |\n|ESM2-650M | 31.87% | 54.63% | 10.66% | 21.01% |\n|ESM-1b | 25.30% | 42.03% | 7.77% | 15.77% |\n\n\n<a name=\"Loc\"></a>\n * <b> Localization (Q10):</b><br/>\n \n| Model | DeepLoc Dataset |\n|--------------------------|:----------------:|\n|**Ankh Large** | 83.01% |\n|Ankh Base | 81.38% |\n|ProtT5-XL-UniRef50 | 82.95% |\n|ESM2-15B | 81.22% |\n|ESM2-3B | 81.22% |\n|ESM2-650M | 82.08% |\n|ESM-1b | 80.51% |\n\n\n<a name=\"RH\"></a>\n * <b> Remote Homology:</b><br/>\n \n| Model | SCOPe (Fold) |\n|--------------------------|:----------------:|\n|Ankh Large | 61.01% |\n|**Ankh Base** | 61.14% |\n|ProtT5-XL-UniRef50 | 59.38% |\n|ESM2-15B | 54.48% |\n|ESM2-3B | 59.24% |\n|ESM2-650M | 51.36% |\n|ESM-1b | 56.93% |\n\n\n<a name=\"Sol\"></a>\n * <b> Solubility:</b><br/>\n \n| Model | Solubility |\n|--------------------------|:----------------:|\n|**Ankh Large** | 76.41% |\n|Ankh Base | 76.36% |\n|ProtT5-XL-UniRef50 | 76.26% |\n|ESM2-15B | 60.52% |\n|ESM2-3B | 74.91% |\n|ESM2-650M | 74.56% |\n|ESM-1b | 74.91% |\n\n\n<a name=\"Flu\"></a>\n * <b> Fluorescence (Spearman Correlation):</b><br/>\n \n| Model | Fluorescence |\n|--------------------------|:----------------:|\n|**Ankh Large** | 0.62 |\n|Ankh Base | 0.62 |\n|ProtT5-XL-UniRef50 | 0.61 |\n|ESM2-15B | 0.56 |\n|ESM-1b | 0.48 |\n|ESM2-650M | 0.48 |\n|ESM2-3B | 0.46 |\n\n\n<a name=\"CATH\"></a>\n * <b> Nearest Neighbor Search using Global Pooling:</b><br/>\n \n| Model | Lookup69K (C) | Lookup69K (A) | Lookup69K (T) | Lookup69K (H) |\n|--------------------------|:----------------:|:----------------:|:----------------:|:----------------:|\n|Ankh Large | 0.83 | 0.72 | 0.60 | 0.70 |\n|**Ankh Base** | 0.85 | 0.77 | 0.63 | 0.72 |\n|ProtT5-XL-UniRef50 | 0.83 | 0.69 | 0.57 | 0.73 |\n|ESM2-15B | 0.78 | 0.63 | 0.52 | 0.67 |\n|ESM2-3B | 0.79 | 0.65 | 0.53 | 0.64 |\n|ESM2-650M | 0.72 | 0.56 | 0.40 | 0.53 |\n|ESM-1b | 0.78 | 0.65 | 0.51 | 0.63 |\n\n\n\n<a name=\"team\"></a>\n## Team\n\n* <b>Technical University of Munich:</b><br/>\n\n| [Ahmed Elnaggar](https://github.com/agemagician) | Burkhard Rost |\n|:------------------------------------------------:|:-------------------------:|\n| <img width=120 src=\"https://github.com/agemagician/Ankh/blob/main/images/AhmedElNaggar.jpg?raw=true\"> | <img width=120 src=\"https://github.com/agemagician/Ankh/blob/main/images/Rost.jpg?raw=true\"> |\n\n\n* <b>Proteinea:</b><br/>\n\n| [Hazem Essam](https://github.com/hazemessamm) | [Wafaa Ashraf](https://github.com/wafaaashraf) | [Walid Moustafa](https://github.com/wmustafaawad) | [Mohamed Elkerdawy](https://github.com/melkerdawy) |\n|:-------------------------:|:-------------------------:|:-------------------------:|:-------------------------:|\n| <img width=120 src=\"https://github.com/agemagician/Ankh/blob/main/images/HazemEssam.jpeg?raw=true\"> | <img width=120 src=\"https://github.com/agemagician/Ankh/blob/main/images/WafaaAshraf.jpeg?raw=true\"> | <img width=120 src=\"https://github.com/agemagician/Ankh/blob/main/images/WalidMoustafa.jpg?raw=true\"> | <img width=120 src=\"https://github.com/agemagician/Ankh/blob/main/images/MohamedElKerdawy.jpeg?raw=true\"> |\n\n\n* <b>University of Columbia:</b><br/>\n\n| [Charlotte Rochereau](https://github.com/crochereau) |\n|:----------------------------------------------------:|\n| <img width=120 src=\"https://github.com/agemagician/Ankh/blob/main/images/CharlotteRochereau.jpg?raw=true\"> |\n\n\n<a name=\"sponsors\"></a>\n## Sponsors\n\n\n| Google Cloud |\n:------------------------------------------------------------------------------------------------------------------------:|\n<img width=120 src=\"https://github.com/agemagician/Ankh/blob/main/images/google_cloud_logo.jpg?raw=true\"> |\n\n\n\n<a name=\"license\"></a>\n## License\nAnkh pretrained models are released under the under terms of the [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International](https://creativecommons.org/licenses/by/4.0/).\n\n<a name=\"community\"></a>\n## Community and Contributions\n\nThe Ankh project is a **open source project** supported by various partner companies and research institutions. We are committed to **share all our pre-trained models and knowledge**. We are more than happy if you could help us on sharing new ptrained models, fixing bugs, proposing new feature, improving our documentation, spreading the word, or support our project.\n\n<a name=\"question\"></a>\n## Have a question?\n\nWe are happy to hear your question in our issues page [Ankh](https://github.com/agemagician/Ankh/issues)! Obviously if you have a private question or want to cooperate with us, you can always **reach out to us directly** via [Hello](mailto:hello@proteinea.com?subject=[GitHub]Ankh). \n\n<a name=\"bug\"></a>\n## Found a bug?\n\nFeel free to **file a new issue** with a respective title and description on the the [Ankh](https://github.com/agemagician/Ankh/issues) repository. If you already found a solution to your problem, **we would love to review your pull request**!.\n\n<a name=\"citation\"></a>\n## \u270f\ufe0f Citation\nIf you use this code or our pretrained models for your publication, please cite the original paper:\n```\n@article{elnaggar2023ankh,\n title={Ankh: Optimized Protein Language Model Unlocks General-Purpose Modelling},\n author={Elnaggar, Ahmed and Essam, Hazem and Salah-Eldin, Wafaa and Moustafa, Walid and Elkerdawy, Mohamed and Rochereau, Charlotte and Rost, Burkhard},\n journal={arXiv preprint arXiv:2301.06568},\n year={2023}\n}\n```\n\n",
"bugtrack_url": null,
"license": "CC-BY-NC-SA-4.0",
"summary": "Optimized Protein Language Model",
"version": "1.10.0",
"project_urls": {
"Homepage": "https://github.com/agemagician/Ankh",
"Repository": "https://github.com/agemagician/Ankh"
},
"split_keywords": [
"ankh"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "74a7ea313596f57c14890f1985ae7a7554698571c239108fa799b352a34441b7",
"md5": "0ef2906a7f56cbdbe093ec6c134b98d8",
"sha256": "bda2909a44aa5e1e7a944e42b2c6e19f8ea433d0aabf899cea4ffe86b6caa8bd"
},
"downloads": -1,
"filename": "ankh-1.10.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "0ef2906a7f56cbdbe093ec6c134b98d8",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.7,<4.0",
"size": 31332,
"upload_time": "2023-05-16T20:31:06",
"upload_time_iso_8601": "2023-05-16T20:31:06.916883Z",
"url": "https://files.pythonhosted.org/packages/74/a7/ea313596f57c14890f1985ae7a7554698571c239108fa799b352a34441b7/ankh-1.10.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "4ec176767b3422b9d2042403dc5e1c441695e9279e804350e333a2c5c1e7b464",
"md5": "a8955b700b660e8db4a23955610455c2",
"sha256": "e26f92d81844a1c5550cda5c5367cb176a4af667cb42a79e17e3ef1a9aefdb97"
},
"downloads": -1,
"filename": "ankh-1.10.0.tar.gz",
"has_sig": false,
"md5_digest": "a8955b700b660e8db4a23955610455c2",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.7,<4.0",
"size": 19827,
"upload_time": "2023-05-16T20:31:08",
"upload_time_iso_8601": "2023-05-16T20:31:08.704032Z",
"url": "https://files.pythonhosted.org/packages/4e/c1/76767b3422b9d2042403dc5e1c441695e9279e804350e333a2c5c1e7b464/ankh-1.10.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-05-16 20:31:08",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "agemagician",
"github_project": "Ankh",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "ankh"
}
JSON
