Sample collection and lake profiling
Tissue samples (~1cm3) were collected from 168 individuals of Suberites diversicolor (Fig 1, Table 1). One lagoon was sampled in Darwin, Australia (DAR), one lagoon and three marine lakes were sampled in Berau (Bay, B.1, B.2 and B.3), and six marine lakes were sampled in Raja Ampat (P.27, P.30, P.32, P.1, P.4 and P.5). Of these locations, nine overlap with the sponge phylogeography study of Becking et al. (2013) and five with the microbial community analysis of Ferreira et al. (2020) (Supplemental Table 1 for corresponding lake codes between the three studies). Samples were collected between 1-5m depth while snorkeling. Some lakes had very low densities of S. diversicolor , therefore sample sizes were lower (see Becking et al., 2013 Table 2 for densities). In the field, tissue samples were immediately preserved in 99% ethanol or RNAlater after excision at 0-4°C (4-8 weeks), and upon returning to the laboratory stored in a -20°C freezer until further use.
Lake characterization was performed concordant with a protocol published in Maas et al. (2018). In brief, lake area (m2) was approximated using Google Earth Pro (v. 7.3.2), maximum depth was measured using a handheld sonar system (Hawkeye), and water parameters (temperature (°C) and salinity (ppt) were measured with an YSI Professional Plus multimeter at 10 locations per lake at 1m intervals from the surface to 5m depth. To define connection to the surrounding sea we measured maximum tidal amplitude simultaneously in the lake and the sea using Hobo water-level loggers. The ratio of maximum tidal amplitude in meters of the lake compared to the sea was used as a proxy to determine as the degree of physical connection between the lake and sea (conform to calculations in Maas et al., 2018).