Molecular techniques offer many benefits for the ecological study of fungi.  In particular, "pooling" (i.e. mixing) many samples of fungal tissue (e.g., mycorrhizas, soil hyphae) followed by molecular analysis offers the potential for much greater replication and statistical power.  Although numerous studies have used a “pooling” approach to characterize fungal communities, little verification has been conducted on how pooling samples affects community descriptions.  To test if pooling can correctly measure fungal species richness and composition, we combined fungi into pools of known species composition and used a set of molecular techniques to see if we could accurately recover these simulated communities.  We used equal amounts of dried and ground sporocarp tissue from fungi previously identified by morphology and molecules to create known combinations of two to 20 species per pool.  DNA from each pool was then extracted and analyzed by Automated Ribosomal Intergenic Spacer Analysis (ARISA), Terminal Restriction Fragment Length Polymorphisms (T-RFLP) , and cloning-sequencing.

Results/Conclusions

Our results indicate that all methods failed to recover the known species number and composition and we found little relationship between the known combinations and the species detected by each method. Our results suggest that pooling can bias descriptions of fungal communities. We will explain how biases in DNA extraction, PCR, or cloning could produce the error and suggest how to improve sampling, at least for ectomycorrhizal fungal studies.