Test Sequences

Table 2.2: Test sets: Number of EST sequences (n) and nucleotides (nt) as raw, trimmed (limited lengths of N-rich regions, poly-A and poly-T sites), and screened (removed ribosomal, chloroplast, and mitochondrial DNA, and remaining sequences shorter than 300 nt) sequences. Sequences were obtained from the library indicated by the ID column. The abbreviation dpi stands for ``days post-inoculation'', indicating mixed plant-microbe cultures.

LIBRARY			RAW		TRIMMED		SCREENED
SPECIES	TISSUE	ID	n	nt	n	nt	n	nt
P. sojae	mycelia	MY	969	527,295	902	510,010	895	506,086
P. sojae	zoospores	ZO	1013	583,520	960	569,576	957	567,976
+G. max	2 dpi	HA	994	577,626	938	563,226	927	556,305
M. truncatula	root hairs	MtRHE	899	539,719	893	536,787	890	534,037
+G. versiforme	10-38 dpi	MHAM	3259	1,785,721	3030	1,735,390	3017	1,725,491
+P. medicaginis	10 dpi	DSIR	2462	1,324,815	2289	1,287,568	2284	1,282,518
M. truncatula	roots	KV0	2718	1,387,832	2550	1,351,137	2492	1,318,131
+S. meliloti	1 dpi	KV1	1125	562,452	1012	537,644	1003	531,813
+S. meliloti	2 dpi	KV2	1960	976,344	1732	926,953	1726	922,433
+S. meliloti	3 dpi	KV3	2375	1,316,430	2217	1,279,691	2173	1,251,795

To diagnose the species of origin for sequences expressed in symbiotic cultures, we collected sequences generated by distinct EST sequencing projects from the GenBank database [13,126,127]. Sequences from pathogenic interactions originated from cultures of a species from the genus Phytophthora with its plant host, such as P. sojae and soybean (G. max) isolated from inoculated hypocotyls two days after infection [93] and P. medicaginis and M. truncatula isolated from infected roots ten days after infection (C. Vance, unpublished). Sequences expressed during mutualistic interactions were obtained from cultures with M. truncatula and mycorrhizal (Glomus versiforme; M.J. Harrison, unpublished) or rhizobacterial (S. meliloti; K. VandenBosch, unpublished) endosymbionts several days after inoculation. Sequences expressed in pure, axenic cultures from P. sojae mycelia and zoospores [93] and from sterile, uninoculated M. truncatula roots [34] provided negative controls.

To maximize the reliability of diagnostic comparisons, we screened test sequences for high quality as for training sequences, and for low similarity to E. coli, chloroplast and mitochondrial genes, and noncoding RNA transcripts (ribosomal and transfer RNAs). Independent BLASTN comparisons identified sequences having very high similarity ($E<10^{-100}$) to vector sequences or moderately high similarity ($E<10^{-20}$) to non-nuclear or non-coding sequences obtained from GenBank. Sequences so identified were withheld from analysis. A summary of test sequences appears in Table 2.2. All test sequences obtained using the procedure described above are available as supplementary material.