Phosformer-ST

Introduction

This repository contains the code to run Phosformer-ST locally described in the manuscript "Phosformer-ST: explainable machine learning uncovers the kinase-substrate interaction landscape". This readme also provides instructions on all dependencies and packages required to run Phosformer-ST in a local environment.

Quick overview of the dependencies

Included in this repository are the following:

• phos-ST_Example_Code.ipynb: ipynb file with example code to run Phosformer-ST

  • modeling_esm.py: Python file that has the architecture of Phosformer-ST

  • configuration_esm.py: Python file that has configuration/parameters of Phosformer-ST

  • tokenization_esm.py: Python file that contains code for the tokenizer

• multitask_MHA_esm2_t30_150M_UR50D_neg_ratio_8+8_shift_30_mask_0.2_2023-03-25_90.txt: this txt file contains a link to the training weights held on the hugging face or zenodo repository

  • See section below (Downloading this repository) to be shown how to download this folder and where to put it

• phosST.yml: This file is used to help create an environment for Phosformer-ST to work

• README.md:

• LICENSE: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International Public License

Installing dependencies with version info

From conda:

Python == 3.9.16

From pip:

For torch/PyTorch

Make sure you go to this website https://pytorch.org/get-started/locally/

Follow along with its recommendation

Installing torch can be the most complex part

Downloading this repository

gh repo clone gravelCompBio/Phosformer-ST 

cd Phosformer-ST 

The following step demonstrates users how to download the training weights

-other repositories were used because the folder's memory size is larger than the allowed space on github

Main option) Hugging Face

Then download the link found in multitask_MHA_esm2_t30_150M_UR50D_neg_ratio_8+8_shift_30_mask_0.2_2023-03-25_90.txt or can be found at this link https://huggingface.co/gravelcompbio/Phosformer-ST_trainging_weights/tree/main

The download link should take to a page that should look like this

Click the download box highlighted in picture above

Alternative option) Zenodo

Then download the link found in multitask_MHA_esm2_t30_150M_UR50D_neg_ratio_8+8_shift_30_mask_0.2_2023-03-25_90.txt or can be found at this link https://zenodo.org/record/8170005

The download link should take to a page that should look like this

Click the download box highlighted in picture above

After picking one of the options above to download the training weights see below

Once downloaded, unizip the folder and place in the Phosformer-ST along with all the other files in this github repository

The final Phosformer-ST directory orinization should have the following files/folder

• file 1 phos-ST_Example_Code.ipynb

• file 2 modeling_esm.py

• file 3 configuration_esm.py

• file 4 tokenization_esm.py

• file 5 multitask_MHA_esm2_t30_150M_UR50D_neg_ratio_8+8_shift_30_mask_0.2_2023-03-25_90.txt

• file 6 phosST.yml

• file 7 Readme.md

• file 8 LICENSE

• folder 1 multitask_MHA_esm2_t30_150M_UR50D_neg_ratio_8+8_shift_30_mask_0.2_2023-03-25_90 (make sure it is unzipped)

🎉 Once you have a folder with the files/folder above you have all the required files to run the model 🎉

Installing dependencies with conda

PICK ONE of the options below

Main Option) Utilizing the PhosformerST.yml file

here is a step-by-step guide to set up the environment with the yml file

Just type these lines of code into the terminal after you download this repository (this assumes you have anaconda already installed)

conda env create -f phosST.yml -n PhosST  

conda deactivate 

conda activate phosST  

Alternative option) Creating this environment without yml file

(This is if torch is not working with your version of cuda or any other problem)

Just type these lines of code into the terminal after you download this repository (this assumes you have anaconda already installed)

conda create -n phosST python=3.9  

conda deactivate 

conda activate phosST  

conda install -c conda-forge jupyterlab 

pip3 install numpy==1.24.3 

pip3 install pandas==2.0.2 

pip3 install matplotlib==3.7.1 

pip3 install scikit-learn==1.2.2 

pip3 install tqdm==4.65.0 

pip3 install fair-esm==2.0.0 

pip3 install transformers==4.31.0 

For torch you will have to download to the torch's specification if you want gpu acceleration from this website https://pytorch.org/get-started/locally/

pip3 install torch torchvision torchaudio 

the terminal line above might look different for you

We provided code to test Phosformer-ST (see section below)

Utilizing the Model with our example code

All the following code examples is done inside of the phos-ST_Example_Code.ipynb file using jupyter lab

Once you have your environment resolved just use jupyter lab to access the example code by typing the command below in your terminal (when you're in the Phosformer-ST folder)

jupyter lab 

Once you open the notebook on your browser, run each cell in the notebook

Testing Phosformer-ST with the example code

There should be a positive control and a negative control example code at the bottom of the phos-ST_Example_Code.ipynb file which can be used to test the model.

Positive Example

# P17612 KAPCA_HUMAN 

kinDomain="FERIKTLGTGSFGRVMLVKHKETGNHYAMKILDKQKVVKLKQIEHTLNEKRILQAVNFPFLVKLEFSFKDNSNLYMVMEYVPGGEMFSHLRRIGRFSEPHARFYAAQIVLTFEYLHSLDLIYRDLKPENLLIDQQGYIQVTDFGFAKRVKGRTWTLCGTPEYLAPEIILSKGYNKAVDWWALGVLIYEMAAGYPPFFADQPIQIYEKIVSGKVRFPSHFSSDLKDLLRNLLQVDLTKRFGNLKNGVNDIKNHKWF" 

# P53602_S96_LARKRRNSRDGDPLP 

substrate="LARKRRNSRDGDPLP" 

  

phosST(kinDomain,substrate).to_csv('PostiveExample.csv') 

Negative Example

# P17612 KAPCA_HUMAN 

kinDomain="FERIKTLGTGSFGRVMLVKHKETGNHYAMKILDKQKVVKLKQIEHTLNEKRILQAVNFPFLVKLEFSFKDNSNLYMVMEYVPGGEMFSHLRRIGRFSEPHARFYAAQIVLTFEYLHSLDLIYRDLKPENLLIDQQGYIQVTDFGFAKRVKGRTWTLCGTPEYLAPEIILSKGYNKAVDWWALGVLIYEMAAGYPPFFADQPIQIYEKIVSGKVRFPSHFSSDLKDLLRNLLQVDLTKRFGNLKNGVNDIKNHKWF" 

# Q01831_T169_PVEIEIETPEQAKTR 

substrate="PVEIEIETPEQAKTR" 

  

phosST(kinDomain,substrate).to_csv('NegitiveExample.csv') 

Both scores should show up in a csv file in the current directory

Inputting your own data for novel predictions

One can simply take the code from above and modify the string variables kinDomain and substrate to make predictions on any given kinase substrate pairs

Formatting of the kinDomain and substrate for input for Phosformer-ST are as follows:

• kinDomain should be a human Serine/Threonine kinase domain (not the full sequence).

• substrate should be a 15mer with the center residue/char being the target Serine or Threonine being phosphorylated

Not following these rules may result in dubious predictions

How to interpret Phosformer-ST's output

This model outputs a prediction score between 1 and 0.

We trained the model to uses a cutoff of 0.5 to distinguish positive and negative predictions

A score of 0.5 or above indicates a positive prediction for peptide substrate phosphorylation by the given kinase

Troubleshooting

If torch is not installing correctly or you do not have a GPU to run Phosformer-ST on, the CPU version of torch is perfectly fine to use

Using the CPU version of torch might increase your run time so for large prediction datasets GPU acceleration is suggested

If you just are here to test if it Phosformer-ST works, the example code should not take too much time to run on the CPU version of torch

Also depending on your GPU the batch_size argument might need to be adjusted

2024-05-17

• if you get an 'EsmTokenizer' object has no attribute 'all_tokens' error when loading the tokenizer
• • Make sure you have version of transformers==4.31.0 installed

The model has been tested on the following computers with the following specifications for trouble shooting proposes

Computer 1

NVIDIA Quadro RTX 5000 (16 GB vRAM)(CUDA Version: 12.1)

Ubuntu 22.04.2 LTS

Intel(R) Xeon(R) Bronze 3204 CPU @ 1.90GHz (6 cores) x (1 thread per core)

64 GB ram

Computer 2

NVIDIA RTX A4000 (16 GB vRAM)(CUDA Version: 12.2)

Ubuntu 20.04.6 LTS

Intel(R) Xeon(R) Bronze 3204 CPU @ 1.90GHz (6 cores) x (1 thread per core)

64 GB ram

Other accessory tools and resources

A webtool for Phosformer-ST can be accessed from: https://phosformer.netlify.app/. A huggingface repository can be downloaded from: https://huggingface.co/gravelcompbio/Phosformer-ST_with_trainging_weights. A huggingface spaces app is available at: https://huggingface.co/spaces/gravelcompbio/Phosformer-ST