Title: A supervised machine learning approach to predict performance and aid decision making of biomaterials design for skin tissue engineering applications

Authors: Aakriti Aggarwal; Pankaj Jain; Saurabh Gupta; Mahesh Kumar Sah

Addresses: Department of Biotechnology, Dr. B.R. Ambedkar National Institute of Technology, Jalandhar, Punjab – 144011, India ' Department of Biomedical Engineering, National Institute of Technology, Raipur, Chhatisgarh – 492010, India ' Department of Biomedical Engineering, National Institute of Technology, Raipur, Chhatisgarh – 492010, India ' Department of Biotechnology, Dr. B.R. Ambedkar National Institute of Technology, Jalandhar, Punjab – 144011, India

Abstract: The physical and chemical interactions among the cells and scaffolds are pivotal for regenerating the desired tissue. The fields of material science and tissue engineering aim to understand this complex behaviour, which can pave new ways for optimising the tissue growth. The present study attempts to predict the in-vitro fibroblast cell growth by modelling the physico-chemical characteristics of the biopolymeric scaffolds through different supervised machine learning strategies for skin tissue engineering application. The chemical nature, porosity, surface roughness, and wettability of the chitosan and gelatine-based scaffolds were used as indicative support; and the cell growth percentage to train various regression models. The random forest classifier provided the specificity, sensitivity, and precision of 88.6%, 99.87%, and 93.75% respectively after hyperparameter tuning. The applicability and efficiency of machine learning for predicting skin tissue engineering outcomes can help in saving time, resources, and human errors while biomaterials designing.

Keywords: biopolymeric scaffolds; cell-scaffold interaction; supervised machine learning; decision making; skin tissue engineering; biomaterial design.

DOI: 10.1504/IJBET.2024.138613

International Journal of Biomedical Engineering and Technology, 2024 Vol.45 No.1, pp.13 - 26

Received: 04 Apr 2023
Accepted: 26 Aug 2023
Published online: 16 May 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article