Understanding the disparity in n-alkane production among angiosperms and gymnosperms from the higher Himalayas: Inferences drawn from a Machine Learning approach

The relative abundances of n-alkanes in plant-derived organic matter serve as reliable biomarkers in paleoecological studies, given that leaf wax characteristics in terrestrial plants are a function of the external environment. Owing to limited phylogenetic and geographic sampling in earlier studies, the relative weight of environmental factors and plant phylogeny on leaf wax production is yet to be well constrained. Here, we try to address the gap in the geographical representation of terrestrial plants and explain their inherent disparity in n-alkane abundance as a function of environment and phylogeny. To do so, we examined 13 plant species, ranging from 500 to 5200 m in altitude, from the northwest to eastern Himalayas, covering three countries, contrasting climate, diverse forest types, over three years. We report: (a) Juniperus (gymnosperm) produces twice the amount of n-alkanes compared to the coexisting angiosperms; (b) gymnosperm n-alkane concentration is seven times higher above the snowline (3000 m) than below due to a paucity of moisture and stronger solar influx; (c) angiosperms n-alkane abundance mimics the precipitation bands in the Himalayas at 2000 m and 4000 m altitude where the n-alkane concentration is three times higher compared to other areas; and (d) the response of plant physiology to climatic forcings is observed in Rhododendron where n-alkane production is connected to the timing of flowering and seed-ripening. We applied a Machine Learning classifier built on Principal Component Analysis and Logistic Regression to disentangle a plant's phylogenetic signal from its external environment, which suggests a 30% higher weight of phylogeny on n-alkane production than the external environment.