Model organisms are typically studied independently of the ecological context in which they evolved. Thus, it is not surprising that many genes in fully sequenced genomes are of unknown function. Some mysterious genes are probably involved in interactions with other organisms. To understand the molecular mechanisms of symbiosis, some investigators and study genes elicited specifically during symbiotic associations, whether mutualistic or pathogenic. This dissertation presents novel methods to identify and analyze symbiotic gene sequences.
The first contribution is a method to determine the originating species of a sequence expressed during interaction. Knowing which genes are expressed by which species can help characterize gene function. Hybridization experiments can make this distinction, but computational techniques can provide the same answer with a known degree of confidence, without consuming costly reagents. The method presented here uses comparative lexical analysis to infer the likely species of origin. Originally used to assess cDNA libraries for contamination, the method has been generalized to determine the species of origin.
The second contribution is a method to estimate diversity in cDNA libraries whose sequencing is ongoing. Estimated transcript diversity provides decision support for when to normalize the rest of the library, yielding the greatest return of distinct transcripts per sample sequenced and increasing the likelihood of sequencing rare transcripts. Non-parametric diversity estimators infer diversity from random samples, approaching true diversity with greater sample size. Validation with distributions having known parameters verifies that the method produces reliable results.
As these methods were developed, they were applied to analyze libraries prepared from pure plant root and fungal cultures, as well as mixed-culture mycorrhizal and pathogenic libraries. The arbuscular mycorrhizal symbiosis is an important and ancient mutualism between fungi of the order Glomales and most vascular plants. A primary insight of the contributions described above is the presence of plant-like transcripts in axenic fungal libraries and many fungal transcripts in a root-hair enriched plant library. The similarity was evident using comparative lexical analysis rather than sequence similarity searching. Interpretations are considered in light of evolutionary history.
NB: Color figures, full content, and supplementary materials are available online at www.santafe.edu/~pth/diss (here):