Chloroplast phylogenomics of Salacia chinensis L. (Celastraceae) with machine learning-assisted insights into anticancer drug discovery
Keywords:
Comparative plastomics; Simple sequence repeats; Phylogenetics; LightGBM; AKT-Scan AI; Tanimoto similarity; Molecular dockingAbstract
The present study reports the first complete chloroplast (Cp) genome of Salacia chinensis L. (Celastraceae), an important medicinal shrub native to Bangladesh, alongside a machine learning-driven exploration of its therapeutic potential. The circular plastome spans 157,454 bp, comprising a large single-copy of 85,757 bp, a small single-copy of 18,451 bp, and two inverted repeats of 26,623 bp each. The Cp genome encodes 127 genes, including 83 protein-coding genes, 36 tRNAs, and eight rRNAs. Comparative plastome analysis indicated a conserved genomic organization with no major structural rearrangements among the closely related members. A total of 95 simple sequence repeats were identified, predominantly mononucleotide motifs (69), suggesting potential markers for genetic diversity studies. Phylogenomic reconstruction confirmed the systematic placement of S. chinensis within Celastraceae. Complementing the genomic insights, a machine learning-guided anticancer drug discovery framework was employed targeting the AKT1 protein (RAC-alpha serine/threonine-protein kinase). A supervised LightGBM model achieved 90.4% accuracy with an AUC of 0.950, enabling the identification of two promising phytochemical leads, Regeol A and Carnaubadiol, exhibiting predicted bioactivities of 51.4% and 68.1%, respectively. Molecular docking analysis demonstrated strong binding affinities of –8.8 kcal/mol and –8.7 kcal/mol for Regeol A and Carnaubadiol, respectively, surpassing the reference drug (–8.0 kcal/mol), while ADMET profiling supported favorable pharmacokinetic properties with minimal toxicity concerns. In addition, we developed AKT-Scan AI (https://aktscanai.streamlit.app), a high-throughput machine learning platform for predicting AKT1-targeted bioactivity and assessing drug-likeness properties. Collectively, this integrative study enriches the genomic understanding of S. chinensis (GenBank Accession: PZ250435.1) and underscores its potential as a promising source of bioactive compounds for targeted therapeutic applications.
Bangladesh J. Plant Taxon. 33(1): 1-19, 2026 (June)
0
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Bangladesh Association of Plant Taxonomists

This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright on any research article is transferred in full to Bangladesh Association of Plant Taxonomists upon publication in Bangladesh Journal of Plant Taxonomy. The copyright transfer includes the right to reproduce and distribute the article in any form of reproduction (printing, electronic media or any other form).

Articles in the Bangladesh Journal of Plant Taxonomy are Open Access articles published under the Creative Commons CC BY License Creative Commons Attribution 4.0 International License.
This license permits use, distribution and reproduction in any medium, provided the original work is properly cited. It is author's responsibility to obtain the permission from appropriate authority if figures are reused from a previously published document.