How many distinct transcripts are expected in a library? As with determining the length of Britain's coastline, the result depends on the scale at which you choose to measure. In the axenic plant library, 726 distinct samples have been found among 899 samples sequenced (at low stringency). Among the pooled fungal libraries, 563 distinct samples are seen among the 701 samples sequenced (at low stringency). Estimated sequence diversity in the plant library is roughly 3200 quasispecies (at low stringency) and about 4800 quasispecies among the pooled fungal libraries.
How does the composition of one fungal library relate to another and to axenic and mixed-culture plant libraries? The fungal libraries share between six and a dozen transcript quasispecies with one another. The three fungal libraries, pooled together are perfectly complementary with pure and mixed-culture plant libraries, sharing no quasispecies.
The pure plant and mixed-culture libraries are inconsistent with respect to the transcripts they share in either state. That is, we might expect the two nodulating root libraries (KVnod and NFnod) or the three axenic root libraries (KV0, NFroot, and MtLong) to cluster together because of similar transcript quasispecies constituents, but this is not the case. The KV0 library overlaps more strongly with KVnod and NFstem libraries, and the NFnod library overlaps with the DSIR and MHAM libraries. This deviation from expectations might have resulted from determining complementarity with non-plant transcripts included. If we were to remove putative microbial sequences from mixed-culture plant libraries and cluster libraries based only those plant transcripts found in each library, the outcome would likely not be the same, perhaps confirming expectations based on biological intuition, or perhaps not. Further inquiry would help to make sense of this curious observation.
More extensive sequencing from any library would both improve the accuracy of gene diversity estimates and increase q, the proportion of transcript quasispecies sequenced. Further sequencing would likely increase the occurrences of related quasispecies between libraries. The high degree of complementarity across fungal libraries may largely be a consequence of small sample sizes or may be an accurate sample of the true diversity. Since axenic plant and fungal libraries do not contain the same quasispecies, it is unlikely that library contamination explains the presence of plant-like transcripts in fungal libraries, and vice versa. This leaves us comfortable in retaining horizontal transfer as a hypothesis to explain the presence of such sequences.
Diversity may be a function of taxonomy or of developmental stage. From spectroscopic assays, we know that fungal spore germ tubes, extraradical hyphae, and intraradical hyphae differ in their lipid and carbohydrate metabolic activies [11,86]. Spore germ tubes are sites of lipid breakdown, rather than synthesis, and of gluconeogenesis. In intraradical hyphae, hexose is transported into the cytoplasm, where lipid synthesis occurs. Extraradical hyphae act as sinks for lipids, where the lipids are sequestered in spores or vesicles. Thus, the fact that the transcript constituents found in extraradical hyphae do not strongly resemble those found in libraries prepared from other developmental stages is not at all surprising. Similar arguments could be made of any specialized tissues, like plant root hairs, whose primary specialized activities are tip elongation and active nutrient transport [49].