What microbial guilds are causing PSF?
We found 66 indicator prokaryotes out of 887 OTUs and 110 indicator fungi out of 1534 OTUs. Guild affiliation was possible for 45 prokaryotes and 33 fungi among our indicator OTUs. We found no significant overrepresentation of fungal pathogens as indicator OTUs of negative plant growth response (Fig. 2, χ2 = 9.79, d.f. = 10, P = 0.46, Fig. 2). Rather, only one of the 33 indicator fungi was a pathogen (negatively affectingBromus ). All the remaining 32 indicator fungi were saprotrophs, and either positively (9) or negatively (3) affected both plants, or negatively affected Bromus only (20). Twelve of the 33 indicator fungi (36%) similarly affected Bromus and Koeleria , while the remainder (64%) had species-specific effects (a necessary condition for microorganisms to drive species-specific PSF). Regarding the prokaryotes, of the 66 indicator OTUs, the majority (37) were associated with improved performance for both Bromus and Koeleria . Conversely, 6 negatively affected Bromus growth and 23 were detrimental to both species. Thus, as opposed to fungi, few indicator prokaryotes (9%, 6 out of 66) had species-specific effects on plants. Indicator prokaryotes were mostly related to N cycling (bacterial and archaeal nitrifiers) or were primary producers (Chloroflexi spp. ). Interestingly, while the vast majority of prokaryotic saprotrophs were associated with positive plant growth responses (18 out of 26 saprotrophic OTUs), fungal saprotrophs tended to be associated with a negative rather than positive plant growth response (23 vs. 9 OTUs, respectively). This contrast between prokaryotes and fungi was statistically significant (χ2 = 8.16,d.f. = 1, P = 0.004).
There was a surprising scarcity of symbiotic indicator OTUs. Out of the 18 arbuscular mycorrhizal fungi and the 16 fungal endophytes (mostly dark septate endophytes) found in our dataset, none were identified as indicator OTU. Similarly, among prokaryotes, we did not find any of the typical plant-growth promoting lineages (e.g., Pseudomonas spp .,Burkholderia spp ., Bacillus spp .) to be associated with consistently improved plant growth in our greenhouse assay, even though they were present in the total microbiome.