MadagascarGeneraAnalysisPipeline

Publications containing phylogenetic reconstructions that included Malagasy species were obtained using two methods: the first was through the G2PMineR automated literature review process (Wojahn et al., 2021) and the second was through manual analysis of the publications underlying the analyses in Buerki et al (2013). The G2PMineR automated literature review process pipeline consisted of two bespoke functions: PhyloPapersSieveR and OutputRestrictoR (see included files). PhyloPapersSieveR took in a csv with one row per species and three columns (family, genus, and estimated number of species), with the columns populated by the Kew list. PhyloPapersSieveR used G2PMineR, rentrez (Winter, 2017), and RISmed (Kovalchik, 2021) internally to query PubMed (Roberts, 2001) and download and mine abstracts. OutputRestrictoR took the output of PhyloPapersSieveR and indicated which publications (1) contained a Malagasy species and (2) were predicted by G2PMineR to contain a dated phylogeny. The publications underlying the analyses in Buerki et al (2013) were then manually downloaded if they were not already included in the output table produced by OutputRestrictoR. The publications were then manually read and the stem and crown ages of any Malagasy species in them were recorded in a spreadsheet. If the stem and crown ages were not available in the text, tables, or supplementary materials of a publication, they were then inferred through direct measurement of the phylogenetic reconstruction using FIJI (Schindelin et al., 2012), with the measurements standardized using the scale provided with the reconstruction. If no such scale was available, then that publication was discarded. A revised species richness for each genus in Madagascar was then added to the final table using the more accurate Catalogue of the Vascular Plants of Madagascar (Phillipson et al., 2006; Phillipson et al., 2022).