Mycobacterium tuberculosis is the causative agent of tuberculosis (TB), which remains a major global health issue, leading to millions of deaths each year [1,2]. The rise of multidrug-resistant (MDR) and extensively drug-resistant (XDR) tuberculosis has greatly reduced the effectiveness of standard treatments [3]. This underlines the necessity to develop innovative drugs that are effective against diverse mechanisms. Natural products have been used as a basic source for the discovery and development of therapeutic medicines. Among these, mushrooms have recently gained considerable scientific interest due to their wide-ranging medicinal properties and biological potential [4]. These fungi are producers of secondary metabolites, many of which have diverse biological activities, including antibacterial, immunomodulatory, anti-inflammatory, and antioxidant activities.

Several studies have underlined the potential of mushroom-derived compounds in fighting mycobacterial infections. Extracts and extracted components from species such as Ganoderma lucidum [5], Pleurotus ostreatus [6], Lentinula edodes [7], and Cordyceps militaris [8] have demonstrated inhibitory effects against Mycobacterium TB in vitro. These bioactive constituents—particularly triterpenoids, polysaccharides, and phenolic compounds—are likely to exert their antimycobacterial effects through diverse mechanisms. These include rupture of the mycobacterial cell wall, inhibition of essential enzyme activity, and modulation of host immunological responses [9,10]. Nonetheless, despite these positive discoveries, a complete understanding of the particular biochemical pathways by which mushroom-derived compounds function against tuberculosis remains inadequate.

Recent developments in computational methodologies have substantially accelerated drug discovery, particularly through the coupling of machine learning (ML) techniques with quantitative structure–activity relationship (QSAR) modeling [11]. ML-based QSAR approaches enable the investigation of large-scale bioactivity datasets, enabling the identification of critical links between chemical structures and biological effects. This, in turn, expedites the discovery of novel pharmacological candidates. One key target in tuberculosis therapy is InhA, an enoyl-acyl carrier protein reductase necessary for the formation of mycolic acids, which are crucial components of the mycobacterial cell wall [12]. Inhibiting InhA activity halts cell wall synthesis and the viability of M. tuberculosis, making it a primary target for innovative anti-TB therapies [13].

In this study, we developed ML-based QSAR models trained on experimentally validated inhibitors of the InhA protein. Using a comprehensive library of mushroom-derived compounds, we conducted virtual screening to identify new scaffolds with predicted anti-TB activity. By leveraging both the chemical diversity of mushrooms and the predictive capabilities of modern machine learning, our approach provides new insights into natural product drug discovery for TB. The results not only support the potential of mushroom metabolites as sources of antimycobacterial agents but also present a scalable computational pipeline for exploring natural products in anti-infective research.