MACHINE LEARNING-ASSISTED RAMAN SPECTROSCOPY FOR MONITORING BIOCATALYTIC NUCLEOSIDE SYNTHESIS IN SUSTAINABLE SOLVENTS

• Grant number: 25/21615-0
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Start date: January 06, 2026
• End date: May 05, 2026
• Field of knowledge: Biological Sciences - Biology
• Principal Investigator: Daniel Sá Freire Lamarca
• Grantee: Mattias Botto Alvarenga Tavares
• Supervisor: Mariano Nicolas Cruz Bournazou
• Host Institution: Faculdade de Ciências Agronômicas (FCA). Universidade Estadual Paulista (UNESP). Campus de Botucatu. Botucatu , SP, Brazil
• Institution abroad: Technical University of Berlin (TU), Germany
• Associated to the scholarship: 25/00427-1 - IDENTIFICATION OF CONTAMINATIONS IN CUPRIAVIDUS NECATOR CULTURES FOR PHB PRODUCTION USING IMAGE ANALYSIS, BP.IC

Abstract

Climate change and the global demand for sustainable biotechnological solutions highlight the urgent need for innovative, scalable, and eco-efficient processes. Among these, the synthesis of nucleosides and their analogues represents a pharmaceutical priority, but conventional chemical methods rely on fossil-derived solvents, energy-intensive conditions, and rare metals, while biocatalytic alternatives face critical challenges such as substrate insolubility in aqueous media. Deep eutectic solvents and their natural variants offer a promising sustainable medium to overcome these solubility limitations; however, they hinder conventional monitoring methods, including HPLC and UV spectroscopy. To address this analytical gap, this project proposes the integration of Raman spectroscopy with machine learning to monitor biocatalytic nucleoside synthesis in DES and NADES. By building spectral databases and training predictive algorithms, the approach aims to enhance real-time monitoring, predict enzyme activity and solubility, and enable process optimization. This framework is expected to contribute to the development of high-throughput, automated, and sustainable bioprocesses, aligned with the United Nations Sustainable Development Goals 9, 12, and 13, while reinforcing the methodological bridge with ongoing research in Brazil on machine learning applied to bioprocess monitoring.