A machine learning-based intrusion detection algorithm for securing bioinformatics pipelines

Osamor, Jude; Yisa, Aliyu; Olanipekun, Febisola; Olowosule, Omotolani; Akerele, Samuel; Anyalechi, Onyekachi; Sadiq, Simbiat; Akerele, Iretioluwa; Palmer, Xavier; Barnett, Michaela

Abstract

Bioinformatics pipelines, which process vast amounts of sensitive biological data, are increasingly targeted by cyberattacks. Traditional security measures often fail to provide adequate protection due to the unique computational and network characteristics of these pipelines. This study proposes a machine learning-based Intrusion Detection System (IDS) tailored specifically for bioinformatics workflows. While the CICIDS2017 dataset serves as the primary benchmark, we augment the study with bioinformatics-specific network traffic to ensure relevance. We compare the performance of four machine learning algorithms Random Forest (RF), Support Vector Machine (SVM), Convolutional Neural Network (CNN), and Gradient Boosting Machine (GBM) and explore hybrid models for enhanced detection. Our findings highlight GBM's superior accuracy (98.3%) while also addressing its computational overhead and susceptibility to adversarial attacks. The study contributes novel insights by integrating real-world bioinformatics traffic data and proposing adaptive security strategies for genomic research environments.

• Presentation Conference Type: Conference Paper (published)
• Conference Name: International Conference on Cyberwarfare and Security
• Start Date: Mar 26, 2025
• End Date: Apr 28, 2025
• Acceptance Date: Mar 3, 2025
• Online Publication Date: Mar 24, 2025
• Publication Date: Mar 24, 2025
• Deposit Date: Apr 5, 2025
• Publicly Available Date: Apr 8, 2025
• Journal: International Conference on Cyber Warfare and Security
• Print ISSN: 2048-9889
• Electronic ISSN: 2048-9870
• Peer Reviewed: Peer Reviewed
• Volume: 20
• Issue: 1
• Pages: 345-353
• Book Title: Proceedings of the 20th International Conference on Cyber Warfare and Security, ICCWS 2025 /
• DOI: https://doi.org/10.34190/iccws.20.1.3363
• Public URL: https://uwe-repository.worktribe.com/output/14269770