INTRODUCTION Potato is an important vegetable and staple food crop around the world because of its high yield per unit area and itsability to act as a nutrient and mineral reservoir. Unfortunately, its productivity was declined due to its susceptibility towards pest and diseases. The single oomycete P. infestans causing devastating yield losses of up to 100% (Nowicki et al., 2012). In this juncture, several strategies like cultural practices, use of chemical fungicides and biocontrol agents have been deployed to manage late blight disease. Of which, the effectiveness of chemical fungicides was lowered due to its negative impact on the environment and the emergence of resistance in the pathogen (Oyesola et al., 2021). Therefore eco-friendly employment of microorganisms in the biological control of pathogens replaces chemical fungicides and serves as a separate line of defense (Shailbala and Kumar 2017). Bacterial endophytes of the genus Bacillus are known to produce antimicrobial biomolecules that are the potent inhibitors of phytopathogens. The bacterial antagonist B. atrophaeus had inhibitory activity against various phytopathogens through their secondary metabolites (Huang et al., 2015; Mu et al., 2020). The in vitro and in vivo screening of putative bioactive chemicals is extremely challenging and time-consuming. Consequently, molecular modelling and docking might simplify the identification of potent anti-oomycete biomolecules and it will aid in characterization of protein target sequencesof P. infestans involved in pathogenicity. In this regard molecular docking of biomolecules 2,4-Di-tert-butylphenol,N-Nitrosopyrrolidine and tetradecanoic acid produced by B. atrophaeusNMB01 against four protein targets of P. infestans were done to discover the biomolecule with anti-oomycete action. MATERIALS AND METHODS Molecular modelling and structure validation of target proteins of P. infestans. The protein targets which are highly essential for the growth, development, survival and pathogenesis of P. infestans were choosen based on literature search. The protein sequences of target proteins were acquired from the Uniprot database. FASTA sequence of the target proteins were subjected to BLAST to compare the protein query percentage in the protein database. Molecular modelling was performed using SWISS-MODEL (Method: Rigid-body assembly) and ROBETTA (Meta server, https://robetta.bakerlab.org/) in accordance with query coverage performance and percentage identity obtained from Blastp. In order to assure that the modeled protein targets have a high level of quality, the Ramachandran plot of the PROCHECK tool on the Structural Analysis and Verification Server (SAVES, Meta server) (https://saves.mbi.ucla.edu/) was used to validate the models of the protein targets. This plot displays which residues are located in regions that are preferred or allowed. The Swiss PDB Viewer (http://www.expasy.org/spdbv/) was utilized for the purpose of energy minimization in modeled proteins as well as loop creation for residues lying in disallowed regions. Ligand preparation. The two dimensional structure of three ligands 2,4-Di-tert-butylphenol, N-Nitrosopyrrolidine and tetradecanoic acid produced by B. atrophaeus NMB01 and the reference ligand molecule mandipropamid as a positive control were retrieved from PubChem database (https://pubchem.ncbi.nlm.nih.gov/) as SDF format. Through Open Babel software, 2D structure was converted to 3D structure in PDB format. Molecular docking. Molecular docking was carried out using the Auto Dock Vina module in PyRx 0.8 (Dallakyan and Olson 2015). Targets were turned into protein macromolecules using PyRx. A 200-step first-order derivative optimization approach using commercial molecular mechanics parametersthe Unified Force Field were used to minimize all ligand structures in order to free up troublesome angles (UFF). The targets binding site pockets were located using the CASTp 3.0 server from the Computed Atlas Topography of Proteins (Tian et al., 2018). When a rigid receptor was used, ligands could yield flexible conformations and orientations with an exhaustiveness value of 8. Docked complex visualization. The BIOVIA Discovery studio client 2021 was updated with the docked conformations of protein-ligand interactions. Observed interactions are registered and then exported as pictures for further analysis. The H-bond surface receptor was utilized so that the ligand binding site could be highlighted more clearly. In order to differentiate between the receptor, the ligand, and the interacting atoms, each one was given a distinct colour. RESULTS AND DISCUSSION Homology and Ab initio modelling. The 3D structure of the target proteins autophagy-related protein 8 and cytochrome c oxidase subunit 1 were modelled using SWISS model server, which had 77.39% and 65.97% sequence identity between template and modelled structure, respectively (Fig. 1). Similarly, for calmodulin and bZIP transcription factor 1, crystal structure was modelled with ROBETTA programme with 65% and 50% confidence score, respectively (Fig. 1). Structure validation. Structural validation of modelled 3D structure of the targetsautophagy-related protein 8, cytochrome c oxidase subunit 1, calmodulin and bZIP transcription factor 1 through Ramachandran plot had 96%, 93.3%, 91% and 84% of amino acid residues in the most favoured region, respectively. Virtual screening and molecular docking. Insilico docking studies between four target proteins of P. infestans (autophagy-related protein 8, cytochrome c oxidase subunit 1, calmodulin and bZIP transcription factor 1) and three ligand molecules (2,4-Di-tert-butylphenol, N-Nitrosopyrrolidine and tetradecanoic acid) revealed that 2,4-Di-tert-butylphenol had the highest docking score of -6.5 kcal/mol which was shown in the heat map (Fig. 2). The binding of 2,4-Di-tert-butylphenol with four target proteins were compared with reference ligand molecule mandipropamid (Fig. 3). The maximum binding energy of 2,4-Di-tert-butylphenol (-6.5 kcal/mol) with the target protein calmodulin could inhibit the communication between plant-microbe interaction and reduce the mRNA levels in the pathogen (Pieterse et al., 1993). Similarly, binding of 2,4-Di-tert-butylphenol with cytochrome c oxidase subunit 1(-6.2 kcal/mol) could impede the supply of ATP essential for pathogenicity by disrupting the electron transport process (Pan et al., 2018). Binding of2,4-Di-tert-butylphenol with bZIP transcription factor 1 and autophagy-related protein 8 could block the appressoria formation by inhibiting the zoospore motility (Blanco et al., 2005) and virulence of P. infestans (Chen et al., 2017). In the previous study, in silico analysis was done to reveal the anti-oomycete nature of 1H-1,2,4-Triazole, 1-octadecanoyl produced during ditrophic interaction of B. subtilis NM261 and P. infestans. Interacting amino acids showing H-bonding, hydrophobic interactions, Van der Waals force, Pi-Alkyl and carbon hydrogen bonds between ligands and target proteins were depicted in Table 1. As the ligand binds to four separate target proteins, resistance in the pathogen is unlikely to emerge because of the various modes of action. Thus, docking results confirmed that 2,4-Di-tert-butylphenolcan be used to manage P. infestans.