Failing to detect other AMF may be ascribed to the short read length with Illumina sequencing (cf. Stockinger et al. 2010). Moreover, Penicillium species (meta-rank 1 here) are common endophytes in plants (Vega et al. 2006), and some species can improve phosphate solubility or produce gibberellic acid to stimulate plant growth (Wakelin et al. 2007; Khan et al. 2008). Fungi may also function as biocontrol agents (e.g., Meira and Candida; Nguyen et al. 2011) or nematode predators (e.g., Dactyllela and Arthrobotrys; Schenck

et al. 1977). Nematodes, common invertebrates in orchids, often cause leaf yellowing and reduce plant vigor (Kuehnle 2006). Such nematophagous fungi may thus play a critical role in controlling nematode infection in orchids. Using symbiotic fungi for controlling disease outbreak or improving the resistance to pathogens has been demonstrated for orchids and

crops (cf. Lee et al. 2009; Wu et al. 2011; Mosquera-Espinosa et al. Doramapimod solubility dmso 2013). Another benefit might be conferred by Sporothrix (meta-rank 2); its abundance is likely associated with the good growth of orchids in Sphagnum moss, a popular potting material in the orchid industry in which Sporothrix is frequently found (Zhang and Andrews 1993; Feeney et al. 2007). Among the well-documented, common pathogenic fungi that infect orchids, Fusarium (meta-rank 4) and Colletotrichum (meta-rank 26) were also detected in this study. Symptoms may be severe enough to impair the growth of Phalaenopsis, KPT-330 mouse e.g., some Fusarium species lead to wilting of orchids (Benyon et al. 1996; Divakaran et al. 2008), and Colletotrichum species cause anthracnose disease (Yang et al. 2011). However, pathogenic species do not always trigger necrotic symptoms because of a lag in symptom expression early during infection (Newton et al. 2010) or the presence of antagonistic species that repress pathogenicity (Schulz and Boyle 2005). Conclusions Metagenomic analysis with NGS techniques provides not only a vast amount of data of barcode sequences, but deep insights into the species composition of a fungal community.

Here, multiple barcodes were used to resolve the taxa within a microbial community; 152 Phospholipase D1 genera (73.8 % OTUs) appeared only in the barcoding with single markers, indicating that no single barcode was able to disclose the diverse microflora comprehensively. Of the six barcodes, ITS1/2, ITS3/4, and nrLSU-U worked the best to decipher the microbiome in Phalaenopsis roots. Our metagenomic analyses suggested that species of the mycorrhizal Trechispora and Mortierella might play some key roles in promoting orchid vigor. Methodological approaches, e.g., in silico simulations on primer preferences, deciphering mock communities with multiple markers, and isolating potentially useful fungi for whole genome sequencing, can be conducted in the future. Acknowledgments This study was financially supported by the National Cheng Kung University and the National Science Council, Taiwan.