Fig. 1: Host shift in D. busckii is mediated by preference for short-chain oligosulfides, specifically DMDS.

a A female D. busckii (Dbus). Phylogenetic relationship between species from three subgenera within the family Drosophilidae. Branch lengths are representative and not to scale. b No-choice bioassay experimental setup and number of eggs laid by each species during 48 h. Significance tested between egg counts of each species. Darkened violin plots indicate significant differences between the number of eggs laid by each species (two-tailed, unpaired t-test with Welch’s correction. Significance p < 0.05). Furthermore, significance was also tested between the number of eggs laid by Dbus against several substrates (one-way ANOVA with multiple comparisons). A filled red box indicates no significant difference in egg numbers when presented with corresponding substrates and compared to water. Dbus and Dmel are represented by green and orange colors, respectively, in either darkened or light color palettes. Note that all substrates were used in a rotting state. c Binary choice assay testing relative oviposition preference (ROP) between rotting orange and a second rotting substrate. Darkened violin plots indicate significant differences between oviposition indices tested between Dmel and Dbus (two-tailed, unpaired t-test with Welch’s correction. Significance p < 0.05). d Binary choice assay testing attraction between rotting oranges and spinach/ mushroom in a BugDorm cage arena (see “Methods” section). Significance was tested using two-tailed, unpaired t-test with Welch’s correction. *p < 0.05. (p < 0.0001 and p = 0.0069 respectively). e Binary choice experiment to test niche separation meditated by a preference for two substrates between two species, where one always was Dmel. The first three rows depict a combined measure of the percent of eggs laid by both species as the eggs could not be morphologically differentiated from each other. The bottom two rows depict the percent eggs laid on each substrate by individual species as it was possible to visually differentiate species-specific eggs. n = 5–9. Error bars represent mean ± SD. f SPME-GC-MS chromatograms of four rotting substrates on a normalized abundance scale. Peaks representing dimethyldisulfide (DMDS) and dimethyltrisulfide (DMTS) are highlighted in green and magenta, respectively. Dimethylsulfide (DMS) was found only in rotting spinach and is highlighted in red. g Binary choice between rotting spinach vs rotting oranges perfumed with DMDS or DMTS (10 μl, 10⁻²in mineral oil each) when tested with Dbus. Significance was tested between control (only rotting orange choice) and treatments using one-way ANOVA followed by multiple comparisons and testing significance between control (column one, only rotting orange choice) and treatments (rotting orange + DMDS/DMTS). ns: p > 0.05, *p < 0.05. The control data (spinach vs orange) was replotted from (c) and reused to compare with the additive effect of short-chain oligosulfides. (column 1 vs 2, p = 0.003; column 2 vs 3, p > 0.9; column 1 vs 3, p = 0.0005). h Binary choice assay testing oviposition preference between DMDS (10 μl, 10⁻² in mineral oil) and mineral oil control in a BugDorm cage arena (see “Methods” section) when tested with Dbus. Significance was tested using two two-tailed, unpaired t-test with Welch’s correction. ns: p > 0.05, *p < 0.05. (DMDS vs control: p = 0.014; DMTS vs control: p = 0.0538; DMDS + DMTS vs control: p < 0.0001) Source data are provided as a source data file.