Most insect communities are composed of evolutionary diverse lineages, but detailed phylogenetic analyses of whole communities are lacking, in particular in species-rich tropical faunas. Likewise, our knowledge of the Tree-of-Life to document life’s evolutionary diversity remains highly incomplete and especially requires the inclusion of unstudied lineages from species-rich ecosystems. Here we present the SITE-100 program, which is an attempt at building the Tree-of-Life from whole-community sampling of high-biodiversity sites around the globe. Combining the local data sets into a global tree produces an increasingly comprehensive tree, while also re-tracing evolutionary history of lineages constituting the local community. Local sets are collected in bulk in standardised passive traps and imaged with a large-scale high-resolution camera, which is followed by a parataxonomy step for the preliminary separation of morphospecies and selection of specimens for phylogenetic analysis. Selected specimens are used for individual DNA extraction and sequencing, usually to sequence mitochondrial genomes. All remaining specimens are bulk extracted and subjected to metabarcoding. Phylogenetic analysis on the mitogenomes produces a reference tree to which short barcode sequences are added in a secondary analysis using phylogenetic placement or backbone constrained
tree searches. However, the approach may be hampered by the fact that (1) mitogenomes are limited in phylogeneticinformativeness, and (2) site-based sampling may produce poor taxon coverage and cause a suite of challenges for phylogenetic inference. To mitigate problems of phylogenetic reconstruction at deep levels, we follow a hierarchical way of gathering molecular information, where nuclear genome and mitogenome data from taxonomically chosen specimens consolidate the base and middle portion of the tree, respectively, and adding species-resolution contributed by DNA barcode data. We posit that site-based sampling, though not meeting the criterion of “taxon-completeness”, has great merits given preliminary studies showing representativeness and evenness of taxa sampled. We therefore argue in favour of site-based sampling as an unorthodox but logistically efficient way to construct large phylogenetic trees.