Evaluation of anaerobic digestion research trends in the last four decades (1980-2025): Substrate digestibility, microbial synergism, and machine learning.

Organic substrates enriched in macromolecules (including carbohydrate, protein, and lipid) are viable sources for energy recovery during anaerobic digestion (AD). Several approaches have been applied to improve AD efficiency. Researchers have introduced new substrates to AD in a continuous flow without considering their full conversion to biomethane (i.e., digestibility). Thus, in this review, all the articles since AD have drawn the attention of researchers (1980-2025) were collected to provide a comprehensive scenario of substrate digestibility. The substrate digestibility was correlated during mono-digestion, co-digestion, pre-treatment, bio-stimulants, and bio-augmentation under varying operational conditions. The microbial dynamics and mechanisms involved in each approach were reviewed, along with the impact of artificial intelligence (AI) on biomethanation to move toward rational and substrate-specific designs in AD. The digestibility was found to be higher for carbohydrates (>50 %), followed by protein-rich (40-50 %) waste in mono-digestion, owing to the intrinsic biodegradation properties and organic loadings. Co-digestion of carbohydrate and protein-rich waste showed digestibility between 20-80 % due to varying volatile fractions and heterogeneous organic compounds in feedstock mixtures. Pre-treatments improved digestibility up to 70 and 60 % for the carbohydrate- and protein-rich waste, respectively. The bio-stimulants (single or combined) promoted digestibility up to 85 % depending on the type of substrate. Bioaugmentation of single or mixed bacterial strains during mixed food waste digestion was mainly reported, in which digestibility was improved to 75 %. Artificial neural networks are among the most effective AI models for biomethane prediction. However, the effect of AI on digestibility is still unclear and needs more extensive studies.