pySAR - Python Sequence Activity Relationship

pySAR is a Python library for analysing Sequence Activity Relationships (SARs)/Sequence Function Relationships (SFRs) of protein sequences.

• The published research article is available here.
• A quick Colab notebook demo of pySAR is available here.
• A Medium article that dives deeper into SARs and the pySAR software itself is available here.

Table of Contents

• Introduction
• Requirements
• Installation
• Usage
• Directories
• Issues
• Contact
• License
• References

Research Article

The research article that accompanied this software is titled: "Machine Learning Based Predictive Model for the Analysis of Sequence Activity Relationships Using Protein Spectra and Protein Descriptors" and was published in the Journal of Biomedical Informatics and is available here [1].

How to cite

Mckenna, A., & Dubey, S. (2022). Machine learning based predictive model for the analysis of sequence activity relationships using protein spectra and protein descriptors. Journal of Biomedical Informatics, 128(104016), 104016. https://doi.org/10.1016/j.jbi.2022.104016

Introduction

pySAR is a Python library for analysing Sequence Activity Relationships (SARs)/Sequence Function Relationships (SFRs) of protein sequences. pySAR offers extensive and verbose functionalities that allow you to numerically encode a dataset of protein sequences using a large abundance of available methodologies and features. The software uses physiochemical and biochemical features from the Amino Acid Index (AAI) database [2], as well as allowing for the calculation of a range of structural, physiochemical and biochemical protein descriptors, via the custom-built protpy package.

After finding the optimal technique and feature set at which to numerically encode your dataset of sequences, pySAR can then be used to build a predictive regression ML model with the training data being that of the encoded protein sequences, and training labels being the in vitro experimentally pre-calculated activity values for each protein sequence. This model maps a set of protein sequences to the sought-after activity value, being able to accurately predict the activity/fitness value of new unseen sequences. The use-case for the software is within the field of Protein Engineering, Directed Evolution and or Drug Discovery, where a user has a set of in vitro experimentally determined activity/fitness values for a library of mutant protein sequences and wants to computationally predict the sought activity value for a selection of mutated unseen sequences, in the aim of finding the best sequence that minimises/maximises their activity value.

In the published research, the sought activity/fitness characteristic is the thermostability of proteins from a recombination library designed from parental cytochrome P450's. This thermostability is measured using the T50 metric (temperature at which 50% of a protein is irreversibly denatured after 10 mins of incubation, ranging from 39.2 to 64.4 degrees C), which we want to maximise [1].

Two additional custom-built softwares were created alongside pySAR - aaindex and protpy. The aaindex software package is used for parsing the amino acid index which is a database of numerical indices representing various physicochemical and biochemical properties of amino acids and pairs of amino acids [2]. protpy is used for calculating a series of protein physiochemical, biochemical and structural protein descriptors. Both of these software packages are integrated into pySAR but can also be used individually for their respective purposes.

A quick Colab notebook demo of pySAR is available here. There is also a Medium article that dives deeper into SARs and the pySAR software itself, available here.

Requirements

• Python >= 3.8
• aaindex >= 1.1.2
• protpy >= 1.2.0
• numpy >= 1.24.2
• pandas >= 1.5.3
• scikit-learn >= 1.2.1
• scipy >= 1.10.1
• tqdm >= 4.65.0
• matplotlib >= 3.6.2
• seaborn >= 0.12.2

Installation

Install the latest version of pySAR via PyPi using pip:

pip3 install pysar --upgrade

Installation from source:

git clone -b master https://github.com/amckenna41/pySAR.git
python3 setup.py install
cd pySAR

Usage

Confile File

pySAR works mainly via JSON configuration files. There are many different customisable parameters for the functionalities in pySAR including the metaparameters of some of the available protein descriptors, all Digital Signal Processing (DSP) parameters in the pyDSP module, the type of regression model to use and parameters specific to the dataset - a description of each parameter is available on the CONFIG.md file.

These config files offer a more straightforward way of making any changes to the pySAR pipeline. The names of All the parameters as listed in the example config files must remain unchanged, only the value of each parameter should be changed, any parameters not being used can be set to null. Additionally, you can pass in the individual parameter names and values to the pySAR and Encoding classes when numerically encoding the protein sequences via kwargs. An example of the config file used in my research project (thermostability.json), with most of the available parameters, can be seen below and in the example config file - CONFIG.md.

{
    "dataset": 
      {
        "dataset": "thermostability.txt",
        "sequence_col": "sequence",
        "activity": "T50"
      },
    "model": 
      {
        "algorithm": "plsregression",
        "parameters": "",
        "test_split": 0.2
      },
    "descriptors":
        {
          "descriptors_csv": "descriptors_thermostability.csv",
          "moreaubroto_autocorrelation":
            {
            "lag":30,
            "properties":["CIDH920105", "BHAR880101", "CHAM820101", "CHAM820102",
              "CHOC760101", "BIGC670101", "CHAM810101", "DAYM780201"],
            "normalize": 1
            },
            ...
      },
    "pyDSP":
      {
        "use_dsp": 1,
        "spectrum": "power",
        "window": {
          "type": "hamming",
          ...
        },
        "filter": {
          "type": null,
          ...
        }
      }
  }

Examples

#import encoding module
from pySAR.encoding import *

'''thermostability.json
{
  "dataset": 
    {
    "dataset": "thermostability.txt",
    "activity": "T50"
    ...
    }
  "model": 
    {
    "algorithm": "randomforest",
    ...
    }
  "pyDSP": 
    {
    "use_dsp": 1,
    "spectrum": "imaginary",
    "window": {
      "type": "blackman"
      }
    }
}
'''
#create instance of Encoding class, using RF algorithm with its default params
encoding = Encoding(config_file='thermostability.json')

#encode sequences using all indices in the AAI if input parameter "aai_indices" is empty/None
aai_encoding = encoding.aai_encoding()

#import encoding module
from pySAR.encoding import *

'''thermostability.json
{
  "dataset": 
    {
    "dataset": "thermostability.txt",
    "activity": "T50"
    ...
    }
  "model": 
    {
    "algorithm": "plsreg",
    "parameters": null
    }
  "pyDSP": 
    {
    "use_dsp": 0,
    ...
    }
}
'''
#create instance of Encoding class, using PLS algorithm with its default params
encoding = Encoding(config_file='thermostability.json')

#encode sequences using 4 indices specified by user, use_dsp = False
aai_encoding = encoding.aai_encoding(aai_indices=["PONP800102","RICJ880102","ROBB760107","KARS160113"])

#import encoding module
from pySAR.encoding import *

'''thermostability.json
{
  "dataset": 
    {
    "dataset": "thermostability.txt",
    "activity": "T50"
    ...
    }
  "model": 
    {
    "algorithm": "adaboost",
    "parameters": [{
      "estimators": 100,
      "learning_rate": 1.5
      ...
    },
  "descriptors": 
  {
    "descriptors_csv": "descriptors_thermostability.csv",
    "moreaubroto_autocorrelation": {
      "lag": 30,
      "properties": ["CIDH920105", "BHAR880101", "CHAM820101", "CHAM820102",
        "CHOC760101", "BIGC670101", "CHAM810101", "DAYM780201"],
      "normalize": 1
    },
    ...
  }
}
'''
#create instance of Encoding class using AdaBoost algorithm, using 100 estimators & a learning rate of 1.5
encoding = Encoding(config_file='thermostability.json')

#building predictive models using all available descriptors, calculating evaluation metrics values for 
# models and storing into desc_results_df DataFrame
desc_results_df = encoding.descriptor_encoding()

#import encoding module
from pySAR.encoding import *

'''thermostability.json
{
  "dataset": 
  {
    "dataset": "thermostability.txt",
    "activity": "T50"
    ...
  }
  "model": 
  {
    "algorithm": "randomforest",
    "parameters": 
      {
      "estimators": 100,
      "learning_rate": 1.5,
      ...
      }
  },
  "descriptors": 
  {
    "descriptors_csv": "descriptors_thermostability.csv",
    "moreaubroto_autocorrelation": {
      "lag": 30,
      "properties": ["CIDH920105", "BHAR880101", "CHAM820101", "CHAM820102",
        "CHOC760101", "BIGC670101", "CHAM810101", "DAYM780201"],
      "normalize": 1
    },
    ...
  },
  "pyDSP": 
  {
    "use_dsp": 0,
    "spectrum": "power",
    "window": ""
    ...
  }
}
'''
#create instance of Encoding class using RF algorithm, using 100 estimators with a learning rate of 1.5 - as listed in config
encoding = Encoding('thermostability.json')

#building predictive models using all available aa_indices + descriptors, calculating evaluation metric values for models and storing into aai_desc_results_df DataFrame
aai_desc_results_df = encoding.aai_descriptor_encoding()

#import pySAR module
from pySAR.pySAR import *

'''thermostability.json
{
  "dataset": 
  {
    "dataset": "thermostability.txt",
    "activity": "T50"
    ...
  },
  "model": 
  {
    "algorithm": "plsregression",
    "parameters": "",
    ...
  },
  "descriptors": 
  {
    "descriptors_csv": "descriptors_thermostability.csv",
    "moreaubroto_autocorrelation": {
      "lag": 30,
      "properties": ["CIDH920105", "BHAR880101", "CHAM820101", "CHAM820102",
        "CHOC760101", "BIGC670101", "CHAM810101", "DAYM780201"],
      "normalize": 1
    },
    ...
  },
  "pyDSP": 
  {
    "use_dsp": 1,
    "spectrum": "power",
    "window": "hamming",
    ...
  }
}
'''
#create instance of PySAR class, inputting path to configuration file
pySAR = PySAR(config_file="thermostability.json")

#encode protein sequences using both the CIDH920105 index + aa_composition descriptor
results_df = pySAR.encode_aai_descriptor(aai_indices="CIDH920105", descriptors="amino_acid_composition")

##########################################################################################
###################################### Parameters ########################################

# AAI Indices: CIDH920105
# Descriptors: amino_acid_composition
# Configuration File: thermostability_config.json
# Dataset: thermostability.txt
# Number of Sequences/Sequence Length: 261 x 466             
# Target Activity: T50
# Algorithm: PLSRegression
# Model Parameters: {'copy': True, 'max_iter': 500, 'n_components': 2, 'scale': True,
#'tol': 1e-06}
# Test Split: 0.2
# Feature Space: (261, 486)

##########################################################################################
######################################## Results #########################################

# R2: 0.6720111107323943
# RMSE: 3.7522525079464457 
# MSE: 14.079398883390391 
# MAE: 3.0713217158459805
# RPD 1.7461053136208489
# Explained Variance 0.6721157080699659

##########################################################################################

#import descriptors class
from pySAR.descriptors import *  

#create instance of descriptors class
desc = Descriptors(config_file="thermostability.json")

#calculate tripeptide composition descriptor
tripeptide_composition = desc.get_tripeptide_composition()

#calculate geary autocorrelation descriptor
geary_autocorrelation = desc.get_geary_autocorrelation()

'''thermostability.json
{
  "dataset": 
  {
    "dataset": "thermostability.txt",
    "activity": "T50"
    ...
  },
  "model": 
  {
    ...
  }
  "descriptors": 
  {
    "descriptors_csv": "descriptors_thermostability.csv",
    "moreaubroto_autocorrelation": {
      "lag": 30,
      "properties": ["CIDH920105", "BHAR880101", "CHAM820101", "CHAM820102",
        "CHOC760101", "BIGC670101", "CHAM810101", "DAYM780201"],
      "normalize": 1
    },
    ...
  },
  "pyDSP": 
  {
    ...
  }
}
'''
#import descriptors class
from pySAR.descriptors import *  

#create instance of descriptors class
desc = Descriptors(config_file="thermostability.json")

#export all descriptors to csv using parameters in config, export=True will export to csv
desc.get_all_descriptors(export=True, descriptors_export_filename="descriptors_thermostability.csv")

from aaindex import aaindex1 

record = aaindex1['CHOP780206'] #get full record
description = aaindex1['CHOP780206'].description #get record's description
refs = aaindex1['CHOP780206'].references     #get record's references 
category = aaindex1['CHOP780206'].category #get record's category
notes = aaindex1['CHOP780206'].notes #get record's notes
correllation_coefficient = aaindex1['CHOP780206'].correllation_coefficient #get record's correllation_coefficient
pmid = aaindex1['CHOP780206'].pmid #get record's pmid
values = aaindex1['CHOP780206'].values  #get amino acid values from record

num_record = aaindex1.num_records()  #get total number of records
record_names = aaindex1.record_names() #get list of all record names
amino_acids = aaindex1.amino_acids() #get list of all canonical amino acids
records = aaindex1.search("hydrophobicity") #get all records with hydrophobicity in their title/description

Directories and Files

• /config - configuration files for the example datasets that pySAR has been tested with, as well as the thermostability.json config file that was used in the research. These config files should be used as a template for future datasets used with pySAR.
• /data - data files used in the research proejct including the thermostability dataset, config file and pre-calculated protein descriptors.
• /docs - documentation for pySAR (pending).
• /example_datasets - example datasets used for the building and testing of pySAR, including the thermostability dataset used in the research. The format of these datasets should be used as a template for future datasets used with pySAR.
• /images - all images used throughout the repo.
• /pySAR - source code for pySAR software.
• /tests - unit and integration tests for pySAR.
• pySAR_research.pdf - published research article.
• pySAR_research.pptx - powerpoint demo of the software development process of pySAR.
• CONFIG.md - example markdown file describing each of the available parameters in the config files.

Issues

Any issues, errors or bugs can be raised via the Issues tab in the repository.

Contact

If you have any questions or comments, please contact amckenna41@qub.ac.uk or raise an issue on the Issues tab.

License

Distributed under the MIT License. See LICENSE for more details.

References

[1]: Mckenna, A., & Dubey, S. (2022). Machine learning based predictive model for the analysis of sequence activity relationships using protein spectra and protein descriptors. Journal of Biomedical Informatics, 128(104016), 104016. https://doi.org/10.1016/j.jbi.2022.104016

[2]: Kawashima, S. and Kanehisa, M., 2000. AAindex: amino acid index database. Nucleic acids research, 28(1), pp.374-374. DOI: 10.1093/nar/27.1.368

[3]: Fontaine NT, Cadet XF, Vetrivel I. Novel Descriptors and Digital Signal Processing- Based Method for Protein Sequence Activity Relationship Study. Int J Mol Sci. 2019 Nov 11;20(22):5640. doi: 10.3390/ijms20225640. PMID: 31718061; PMCID: PMC6888668.

[4]: Cadet, F., Fontaine, N., Li, G. et al. A machine learning approach for reliable prediction of amino acid interactions and its application in the directed evolution of enantioselective enzymes. Sci Rep 8, 16757 (2018).

[5]: Lutz S. Beyond directed evolution--semi-rational protein engineering and design. Curr Opin Biotechnol. 2010 Dec;21(6):734-43. doi: 10.1016/j.copbio.2010.08.011. Epub 2010 Sep 24. PMID: 20869867; PMCID: PMC2982887.

[6]: Yang, K.K., Wu, Z. & Arnold, F.H. Machine-learning-guided directed evolution for protein engineering. Nat Methods 16, 687–694 (2019). https://doi.org/10.1038/s41592-019-0496-6

[7]: Yuting Xu, Deeptak Verma, Robert P. Sheridan, Andy Liaw, Junshui Ma, Nicholas M. Marshall, John McIntosh, Edward C. Sherer, Vladimir Svetnik, and Jennifer M. Johnston Journal of Chemical Information and Modeling 2020 60 (6), 2773-2790 DOI: 10.1021/acs.jcim.0c00073

[8]: Medina-Ortiz, D., Contreras, S., Amado-Hinojosa, J., Torres-Almonacid, J., Asenjo, J. A., Navarrete, M., & Olivera-Nappa, Á. (2020). Combination of digital signal processing and assembled predictive models facilitates the rational design of proteins. ArXiv [Cs.CE].

Back to top