To evaluate the predictions made by comparative hexamer analysis with sequences whose origin and function have been characterized, we studied a set of 10 sequences cloned from G. intraradices, 4 sequences from G. versiforme, and 12 sequences from M. truncatula. Some of these sequences have been previously analyzed, in Section 2.4, Table 2.3. Others were recently accessioned to the GenBank database; some are shorter than the 300 nt sequences studied in Chapter 2.
Several fungal chitin synthases were studied in detail. Chitin is a polysaccharide that constitutes a major structural component of fungal cell walls, and is thus essential to fungal growth and development [63,103]. Several variants of chitin synthase have been sequenced, and constitute the majority of protein-coding sequences (15 of 27, as of April 2001 AD) from Glomus spp. in GenBank.
Several plant chitinases were chosen for analysis because of their role in plant defense and mycorrhizal colonization. Plants deploy chitinases as a means of defense against pathogens [15,63], but one class of chitinase is specific to the mycorrhizal association [103]. Because of the role of chitinases in plant defense and pathogen infection, rapid evolution has been detected among plant chitinases [15].
Where available, other plant and fungal genes thought to be involved in the mycorrhizal association were chosen for analysis as validation sequences, including several plant and fungal phosphate transporters.
The validation sequences chosen for analysis, and partial transcripts of the same genes, were withheld from training sets and subjected to the same trimming criteria as described below in Section 4.2.2. Sequence accession numbers and the source molecule type of these 26 sequences, indicated by the ``mRNA?'' column, are shown in Table 4.3.