The Effect of Different Methacrylation Amounts on Physical Properties of Gelatin Methacryloyl Biomaterials: Machine Learning Approach

Abstract

The rational design process for biomaterials is time-consuming. Machine learning (ML) is an efficient approach for reducing material synthesis and experimentation in terms of cost and time. Among the emerging biopolymers for tissue engineering applications, methacrylic anhydride (MA)-functionalized gelatin (GelMA), which was chosen as the model biomaterial for this study, has assumed a promising role owing to its excellent tunable properties and biocompatibility. The ML approach was used to determine the efficiency of the MA amounts selected for GelMA synthesis. In addition, the effect of different methacrylation amounts on the molecular structure of GelMA was indicated in terms of its physical properties. This modeling was performed to generate predictions based on 20 mL of MA. The prediction output was obtained as a result of four data models from the 20 mL MA column. First, data were collected with experimental applications for swelling and degradation ratios, and then the data processing phase was applied. The most suitable ML model, decision tree regression, was selected, and the results were interpreted graphically. The experimental results were compared with the ML results, and the efficiency of ML is shown in detail. The Mean Squared Error (MSE) value for degradation was calculated as 10.16, with a Root Mean Squared Error (RMSE) of 3.1885, Mean Absolute Error (MAE) of 2.6667, and Mean Absolute Percentage Error (MAPE) of 14.66%. For swelling, the MSE value was calculated to be 1821.25, with an RMSE of 3.1885, MAE of 2.6667, and MAPE of 14.66%. In future studies, it is anticipated that the performance of the model will improve with the expansion of the experimental dataset for swelling measurements.

Keywords

• methacrylic anhydride
• Gelatin
• GelMA
• physical properties
• data analysis
• Machine learning

References

1. 1. Meyer TA, Ramirez C, Tamasi MJ, Gormley AJ. A user’s guide to machine learning for polymeric biomaterials. ACS Polym Au [Internet]. 2023 Apr 12;3(2):141–57. Available from: .
2. 2. Tutar R, Koken SY, Tuncaboylu DC, Çelebi-Saltik B, Özeroğlu C. In situ formation of biocompatible and ductile protein-based hydrogels via Michael addition reaction and visible light crosslinking. New J Chem [Internet]. 2023;47(22):10759–69. Available from: .
3. 3. Basu B, Gowtham NH, Xiao Y, Kalidindi SR, Leong KW. Biomaterialomics: Data science-driven pathways to develop fourth-generation biomaterials. Acta Biomater [Internet]. 2022 Apr;143:1–25. Available from: .
4. 4. Greener JG, Kandathil SM, Moffat L, Jones DT. A guide to machine learning for biologists. Nat Rev Mol Cell Biol [Internet]. 2022 Jan 13;23(1):40–55. Available from: .
5. 5. Inza I, Calvo B, Armañanzas R, Bengoetxea E, Larrañaga P, Lozano JA. Machine learning: An indispensable tool in bioinformatics. In 2010. p. 25–48. Available from: .
6. 6. Peng GCY, Alber M, Buganza Tepole A, Cannon WR, De S, Dura-Bernal S, et al. Multiscale modeling meets machine learning: What can we learn? Arch Comput Methods Eng [Internet]. 2021 May 17;28(3):1017–37. Available from: .
7. 7. Castelli V, Cover TM. On the exponential value of labeled samples. Pattern Recognit Lett [Internet]. 1995 Jan 1;16(1):105–11. Available from: .
8. 8. Reddy YCAP, Viswanath P, Reddy BE. Semi-supervised learning: a brief review. Int J Eng &Technology [Internet]. 2018;7(1):81–5. Available from: .

9. 9. Li Y. Deep Reinforcement Learning: An Overview [Internet]. 2017. 85 p. Available from: .
10. 10. Ayodele TO. Types of Machine Learning Algorithms. In: New Advances in Machine Learning [Internet]. InTech; 2010. Available from: .
11. 11. Mahesh B. Machine learning algorithms - A review. Int J Sci Res [Internet]. 2020 Jan 5;9(1):381–6. Available from: .
12. 12. Aery MK, Ram C. A review on machine learning: Trends and future prospects. An Int J Eng Sci [Internet]. 2017;25:89–96. Available from: .
13. 13. Pathak S, Mishra I, Swetapadma A. An assessment of decision tree based classification and regression algorithms. In: Proceedings of the International Conference on Inventive Computation Technologies (ICICT-2018) [Internet]. 2018. p. 92–5. Available from: .
14. 14. Myles AJ, Feudale RN, Liu Y, Woody NA, Brown SD. An introduction to decision tree modeling. J Chemom [Internet]. 2004 Jun 4;18(6):275–85. Available from: .
15. 15. Süren SM, Tutar R, Özeroğlu C, Karakuş S. Versatile multi-network hydrogel of acrylamide, sodium vinyl sulfonate, and N,N′-methylene bisacrylamide: A sustainable solution for paracetamol removal and swelling behavior. J Polym Environ [Internet]. 2024 Jan 20;32(1):164–81. Available from: .
16. 16. Tavafoghi M, Sheikhi A, Tutar R, Jahangiry J, Baidya A, Haghniaz R, et al. Engineering tough, injectable, naturally derived, bioadhesive composite hydrogels. Adv Healthc Mater [Internet]. 2020 May 24;9(10):1901722. Available from: .
17. 17. Van Den Bulcke AI, Bogdanov B, De Rooze N, Schacht EH, Cornelissen M, Berghmans H. Structural and rheological properties of methacrylamide modified gelatin hydrogels. Biomacromolecules [Internet]. 2000 Mar 14;1(1):31–8. Available from: .
18. 18. Nichol JW, Koshy ST, Bae H, Hwang CM, Yamanlar S, Khademhosseini A. Cell-laden microengineered gelatin methacrylate hydrogels. Biomaterials [Internet]. 2010 Jul;31(21):5536–44. Available from: .
19. 19. He J, Sun Y, Gao Q, He C, Yao K, Wang T, et al. Gelatin methacryloyl hydrogel, from standardization, performance, to biomedical application. Adv Healthc Mater [Internet]. 2023 Sep 15;12(23):2300395. Available from: .
20. 20. Noshadi I, Hong S, Sullivan KE, Shirzaei Sani E, Portillo-Lara R, Tamayol A, et al. In vitro and in vivo analysis of visible light crosslinkable gelatin methacryloyl (GelMA) hydrogels. Biomater Sci [Internet]. 2017;5(10):2093–105. Available from: .
21. 21. O’Connell CD, Zhang B, Onofrillo C, Duchi S, Blanchard R, Quigley A, et al. Tailoring the mechanical properties of gelatin methacryloyl hydrogels through manipulation of the photocrosslinking conditions. Soft Matter [Internet]. 2018;14(11):2142–51. Available from: .
22. 22. Karaoglu IC, Kebabci AO, Kizilel S. Optimization of gelatin methacryloyl hydrogel properties through an artificial neural network model. ACS Appl Mater Interfaces [Internet]. 2023 Sep 27;15(38):44796–808. Available from: .
23. 23. Van Vlierberghe S, Dubruel P, Schacht E. Effect of cryogenic treatment on the rheological properties of gelatin hydrogels. J Bioact Compat Polym [Internet]. 2010 Sep 4;25(5):498–512. Available from: .
24. 24. Chen Y, Lin R, Qi H, Yang Y, Bae H, Melero‐Martin JM, et al. Functional human vascular network generated in photocrosslinkable gelatin methacrylate hydrogels. Adv Funct Mater [Internet]. 2012 May 23;22(10):2027–39. Available from: .
25. 25. Tutar R, Yüce-Erarslan E, İzbudak B, Bal-Öztürk A. Photocurable silk fibroin-based tissue sealants with enhanced adhesive properties for the treatment of corneal perforations. J Mater Chem B [Internet]. 2022;10(15):2912–25. Available from: .
26. 26. Rahali K, Ben Messaoud G, Kahn C, Sanchez-Gonzalez L, Kaci M, Cleymand F, et al. Synthesis and characterization of nanofunctionalized gelatin methacrylate hydrogels. Int J Mol Sci [Internet]. 2017 Dec 10;18(12):2675. Available from: .
27. 27. Claaßen C, Claaßen MH, Truffault V, Sewald L, Tovar GEM, Borchers K, et al. Quantification of substitution of gelatin methacryloyl: Best practice and current pitfalls. Biomacromolecules [Internet]. 2018 Jan 8;19(1):42–52. Available from: .
28. 28. Lee Y, Lee JM, Bae P, Chung IY, Chung BH, Chung BG. Photo‐crosslinkable hydrogel‐based 3D microfluidic culture device. Electrophoresis [Internet]. 2015 Apr 24;36(7–8):994–1001. Available from: .
29. 29. Jamal P, Ali M, Faraj RH, Ali PJM, Faraj RH. Data normalization and standardization: A technical report. Mach Learn Tech Reports [Internet]. 2014;1(1):1–6. Available from: .
30. 30. Tso GKF, Yau KKW. Predicting electricity energy consumption: A comparison of regression analysis, decision tree and neural networks. Energy [Internet]. 2007 Sep;32(9):1761–8. Available from: .
31. 31. Yang L, Shami A. On hyperparameter optimization of machine learning algorithms: Theory and practice. Neurocomputing [Internet]. 2020 Nov;415:295–316. Available from: .
32. 32. Hodson TO. Root-mean-square error (RMSE) or mean absolute error (MAE): when to use them or not. Geosci Model Dev [Internet]. 2022 Jul 19;15(14):5481–7. Available from: .
33. 33. Polat K, Güneş S. Automatic determination of diseases related to lymph system from lymphography data using principles component analysis (PCA), fuzzy weighting pre-processing and ANFIS. Expert Syst Appl [Internet]. 2007 Oct;33(3):636–41. Available from: .
34. 34. İnal M. Determination of dielectric properties of insulator materials by means of ANFIS: A comparative study. J Mater Process Technol [Internet]. 2008 Jan;195(1–3):34–43. Available from: .
35. 35. Amid S, Mesri Gundoshmian T. Prediction of output energies for broiler production using linear regression, ANN (MLP, RBF), and ANFIS models. Environ Prog Sustain Energy [Internet]. 2017 Mar 7;36(2):577–85. Available from: .
36. 36. Willmott CJ, Matsuura K. Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance. Clim Res [Internet]. 2005 Dec 19;30(1):79–82. Available from: .
37. 37. Loh W. Classification and regression trees. WIREs Data Min Knowl Discov [Internet]. 2011 Jan 6;1(1):14–23. Available from: .