Analysis of the Natural Foci of Tick-Borne Pathogens in Guangxi Based on Multi-Source Data and Machine Learning

摘要

Located on the northeastern edge of the Indochinese Peninsula, Guangxi serves as a critical ecological hub. Its humid subtropical climate and complex habitats provide a natural reservoir for the cross-host transmission of tick-borne pathogens. This study integrated multi-source data on 77 tick batches and their habitats across 14 prefectures in Guangxi between 2019 and 2024 to develop a machine-learning-driven framework for reservoir analysis. Random forest models revealed that daily average precipitation (importance 0.21) and vegetation type are the core drivers of tick species distribution (P ≈ 0). The micro-fan-headed tick (Ixodes microfanus) is the dominant species (accounting for 61.3%) due to its broad adaptability (temperature and humidity thresholds of 18–28°C and 65–85%, respectively). While host-specific species such as the Tamoc tick exhibit localized distributions, the Java flower tick shows strong host specificity for the Malayan pangolin. A heatmap of the host transmission network reveals cross-species transmission pathways of the long-horned tick between wild boars and domestic dogs. The XGBoost model (AUC = 0.97) identified a temperature threshold (>24°C) two weeks before onset as a key trigger for scrub typhus outbreaks (OR = 4.8), which increases infection risk through a “temperature-larval activity” chain reaction. This study provides a data-driven paradigm for the precise prevention and control of tick-borne diseases in border regions. It enriches the understanding of tick-borne infectious diseases in natural foci within the tropical-subtropical transition zone.