An integrated 3-M workflow for accelerated annotation of natural products: Flavonoids in Daemonorops draco as a case study
Wenxiang Fan,Ziwei Li,Longchan Liu,Yu Wang,Kaixian Chen,Linnan Li,Zhengtao Wang,Li Yang
DOI: https://doi.org/10.1016/j.talanta.2024.126921
IF: 6.1
2025-01-01
Talanta
Abstract:Efficient annotation and dereplication of metabolites, particularly those from resource-endangered plants lacking reference standards, is crucial for natural products development. Advanced techniques like high resolution mass spectrometry (LC-HRMS) have significantly enhanced metabolite characterization. However, challenges such as redundant spectral data, limited reference databases, and inferior dereplication capacity hinder its broad applicability. In this study, we propose an integrated annotation strategy utilizing various computational tools, including mass defect filters (MDF), molecular fingerprints, and molecular networks (3-M strategy). We demonstrate this approach using Daemonorops draco (D. draco), a renowned yet resource-endangered natural product rich in functional flavonoids. By applying pre-defined flavonoids MDF windows, the MS1 peaks reduced by 85 % (from 10,043 to 1,585) in positive mode. Subsequent de novo molecular formula annotation and molecular fingerprint-based structure elucidation were automatically performed using the SIRIUS machine learning platform. Additionally, two complementary cluster tools were incorporated, including feature-based molecular network (FBMN) and t-distributed stochastic neighbor embedding (t-SNE) molecular network, to efficiently dereplicate metabolites and discover novel flavonoids in D. draco. Totally, 108 flavonoids (containing flavones, flavanes, flavanones, chalcones, chalcanes, dihydrochalcones, anthocyanins, homoisoflavanes, homoisoflavanones, and isoflavones), 18 flavone derivatives, and 54 flavone oligomers were identified. Among them, 25 compounds were firstly reported in D. draco. This 3-M workflow shed light on the composition of D. draco and validate the effectiveness of our approach, which facilitated the rapid annotation and screening of subclass metabolites in complex natural products.