Identification of Diseases caused by non-Synonymous Single Nucleotide Polymorphism using Machine Learning Algorithms - Abstract

Aug 4, 2026

Upload by Levi

Vaccine development relies on comprehensive analysis of disease genetics, yet experimentally testing every known gene variant is costly and time-consuming. This study leverages multiple computational machine learning approaches to identify and predict single nucleotide polymorphisms, with emphasis on non-synonymous SNPs. Models including SVM, Random Forest, Logistic Regression, KNN, Naïve Bayes, Decision Trees, Neural Networks, and Gradient Boosting are compared for disease-related variant characterization. Results show Gradient Boosting achieving about 70% accuracy.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages14

Reading Time35 min

AccessFree to download

Introduction

Background

Materials and Methods

Results and Discussion

Conclusion

Keywords: Single Nucleotide Polymorphism, in-silico, Machine Learning, Bioinformatics Journal Info: Submitted: November 24, 2024 Accepted: December 19, 2024 Published: December 31, 2024 Identiﬁcation of Diseases caused by non-Synonymous Single Nucleotide Polymorphism using Machine Learning Algorithms Shaheen 2,4 , Saman Safdar 1 1 Department of Computer Science, COMSATS University Islamabad, Lahore Campus, Lahore, Pakistan; 2 School of Systems and Technology, University of Management and Technology, Lahore, Pakistan; 3 Department of Biochemistry, CMH Medical College, Lahore, Pakistan; 4 School of Arts and Digital Industries, University of Roehampton, London, United Kingdom Abstract The production of vaccines for diseases depends entirely on its analysis. However, to test every disease extensively is costly as it would involve the investigation of every known gene related to a disease. This issue is further elevated when different variations of diseases are considered. As such the use of different computational methods are considered to tackle this issue. This research makes use of different machine learning algorithms in the identiﬁcation and prediction of Single Nucleotide Polymorphism. This research presents that Gradient Boosting algorithm performs better in comparison to other algorithms in genic variation predictions with an accuracy of 70% . *Correspondence author email address: [junaidanjum@cuilahore.edu.pk](junaidanjum@cuilahore.edu.pk)[ ](junaidanjum@cuilahore.edu.pk)DOI: 10.21015/vtse.v12i4 .1984 1 Introduction The investigation of genetic variations is currently being done through different practices such as Gene Panel which basically provides identiﬁcation of variants in more than one gene [1] . Through this test it is possible to pinpoint a disease and its different symptoms. However, this process can take a relatively long time. Because of this, recently, researchers have shifted their focus towards the identiﬁcation of gene variants using computational methods or more speciﬁcally through the use of Machine Learning (ML) [2][3][4][5] . Over several years, different ML algorithms have been developed that aim to speed up the investigation of variants in genes. The main goal behind the use of ML is to speed up the identiﬁcation process as well as lower the cost involved in the testing process. This speed-up process of identiﬁcation would allow researchers to gather information about the different variants of a gene and thus help in diagnosing and creating a medicine against the This work is licensed under a Creative Commons Attribution 3.0 License. VFAST Transactions on Software Engineering Volume 12, Issue 4, 2024 investigated variant. This research focuses on the identiﬁcation of single nucleotide polymorphism (SNP) or, more speciﬁcally, its sub-type of non-synonymous SNP (nsSNP), also known as single amino acid polymorphism (SAP) [6] . Using the concept of ML, this research provides the identiﬁcation of different diseases that are caused by substitutions, insertions, or deletions in an amino acid sequence of a particular protein. The identiﬁcation is done using different ML algorithms which comprised of Support Vector Machine (SVM), Random Forest Classiﬁer, Logistic Regression, K-Nearest Neighbor, Naïve Bayes, Decision Trees, Neural Networks and Gradient Boosting Classiﬁer. All these ML algorithms were compared with each other to determine which performed best in the prediction of different diseases caused by nsSNPs. The novelty of this research work is that it provides the use of in-silico methods to extract features from different diseases in human beings that are caused by the modiﬁcation of nsSNPs in proteins. Using the extracted features, different ML models are employed that can be used to accurately identify diseases. This research work also provides a comprehensive review in regard to the selected ML models and their applications. The article is divided into d

FAQs

What problem does this research address in disease identification?

Comprehensive experimental testing across all known disease-linked genes is expensive and slow. The work aims to speed up identification using computational methods and machine learning.

Which genetic variant type is the focus of the study?

The study targets single nucleotide polymorphisms, specifically non-synonymous SNPs (nsSNP), also associated with single amino acid polymorphism effects.

Which machine learning approach performed best and what accuracy was reported?

Gradient Boosting performed better than other compared algorithms, achieving an accuracy of about 70% for genic variation predictions.

Aug 4, 2026

Upload by Levi