Dissecting Trichoderma antagonism: Role of strain identity, volatiles, biomass, and morphology in suppressing cacao pathogens

Cacao farming worldwide suffers from damaging fungal infections, including those caused by Colletotrichum gloeosporioides and Pestalotiopsis spp., which prioritizes the development of sustainable control solutions. This study evaluated the in vitro antagonistic potential of three Trichoderma strains (T. virens 11C-65-1, T. virens 29-8, Trichoderma spp. RC) against six cacao pathogen isolates from Ghana. Dual culture assays, detailed morphological analysis of both antagonist and pathogen, volatile organic compound (VOC) assays (comparing standard plug vs. pre-grown biomass methods), and the effect of UVC pretreatment were investigated. Multivariate statistics and machine learning were employed to analyze interaction dynamics and predict outcomes. All tested Trichoderma strains significantly inhibited pathogen growth, with efficacy varying notably: 11C-65-1 > RC > 29-8. Significant morphological changes were observed in both interacting fungi. Machine learning models predicted pathogen colony size with high accuracy (test set R² up to 0.94), identifying the specific Trichoderma-pathogen pair identity and Trichoderma circularity as the most crucial predictors. VOCs contributed to inhibition, and using a larger Trichoderma biomass drastically increased antagonistic effects, likely through combined VOC action and physical interaction. UVC pretreatment induced statistically significant but minimal morphological changes under the tested conditions. Multivariate analyses linked Trichoderma strain identity strongly with the resulting circularity of both antagonist and pathogen. Key takeaways from this research are the pronounced strain specificity governing Trichoderma antagonism, the indication of multiple active mechanisms, the relevance of morphology in these fungal battles, and the identification of T. virens 11C-65-1 as a strong prospect for biocontrol application against cacao pathogens, offering a data-driven approach for selecting effective biocontrol agents in other agricultural systems.