This Article Abstract Full Text Services Email this article to a colleague Alert me to new issues of the journal Request Copyright Permission Citing Articles Citing Articles via CrossRef

Manning et al. 10.1073/pnas.0710834105.

Supporting Information

Files in this Data Supplement:

Fig. 5. Circular map of the Escherichia coli Sakai complete genome and comparisons with the spinach outbreak strain partial genome and the EDL-933 complete genome. The outer two circles show Sakai protein coding genes colored by Clusters of Orthologous Groups (COGs) of proteins (1). Genes on the forward strand are shown by the outside circle, and genes on the reverse strand are shown by the inside circle. In circles 3 and 4, Sakai genes conserved in EDL-933 are in blue; nonconserved genes are in gray. In circles 5 and 6, Sakai genes conserved in the spinach strain are in gold; nonconserved genes are in gray. Circles 7 and 8 show Sakai genes containing SNPs in EDL-933. Circles 9 and 10 show Sakai genes containing SNPs in the spinach strain. These SNP harboring genes are colored by the number of SNPs: 1-5 SNPs in green; 6-10 SNPs in blue; 11-20 SNPs in orange; >20 SNPs in red. The number of highly conserved genes (n = 2,741) is highlighted among three O157 genomes. The Sakai and EDL-933 genomes are more similar to each other in gene content and nucleotide sequence identity (3.2%) than to the clade 8 spinach outbreak strain (10.65% or 10.7%).

1. Huson DH (1998) SplitsTree: Analyzing and visualizing evolutionary data. Bioinformatics 14:68-73.

Fig. 6. Year by year changes in the number of reported cases of Escherichia coli O157:H7 in Michigan (n = 444). The decrease in the annual number of cases in Michigan from 2002 follows the national trend in E. coli O157:H7 disease (dotted line). The percentage of strains representing clade 8 has increased in frequency over time (purple line), whereas clade 2 frequency has decreased (blue line).

Table 2. Distribution and frequency of SNP genotypes (SGs) among 528 Escherichia coli O157 strains and close relatives

Note: Strain isolation dates are represented by commas for SGs with fewer than two strains and as a range for categories with more strains and those with an unknown collection date.

*SG-35 contains seven strains including (β-glucuronidase-positive, GUD⁺; sorbitol-negative, Sor^-) strains that are O157:H7. SG-36 contains six strains isolated in Germany that are GUD⁺/Sor⁺ and have serotype O157:H^-. SG-37 strain represents a nontypeable (NT) serotype (O antigen) isolated from a healthy marmoset.

†Strains are O55:H7 serotypes and represent the evolutionarily derived lineages (GUD^-/Sor^-)

Table 3. Crude associations between patient characteristics and infection with Escherichia coli O157 strains (n = 333) of different clades

	Clade 8* (n = 63)				Clade 2* (n = 154)				Clade 7* (n = 31)
Characteristic†	n	%	OR (95% CI)	P	n	(%)	OR (95% CI)	P	n	%	OR (95% CI)	P
Bloody diarrhea
No (n = 57)	8	14	1.0		25	(43)	1.0		16	28
Yes (n = 234)	55	24	1.8 (0.84, 4.21)	0.11	129	(55)	1.6 (0.88, 2.81)	0.13	15	6	0.2 (0.08, 0.38)	<0.0001
Non-bloody diarrhea
No (n = 112)	23	21	1.0		64	(57)	1.0		13	12
Yes (n = 179)	40	22	1.1 (0.62, 1.98)	0.71	90	(50)	0.8 (0.47, 1.22)	0.25	18	10	0.9 (0.40, 1.81)	0.68
Abdominal pain
No (n = 52)	7	13	1.0		26	(50)	1.0		8	15	1.0
Yes (n = 239)	56	23	2.0 (0.84, 4.61)	0.10	128	(54)	1.2 (0.63, 2.10)	0.64	23	10	0.6 (0.25, 1.39)	0.24
Body aches
No (n = 244)	53	22	1.0		126	(52)	1.0		26	11	1.0
Yes (n = 47)	10	21	1.0 (0.45, 2.09)	0.95	28	(60)	1.4 (0.73, 2.60)	0.32	5	11	1.0 (0.36, 2.75)	1.0
HUS
No (n = 281)	56	20	1.0		151	(54)	1.0		31	11	NA
Yes (n = 10)	7	70	9.4 (2.35, 37.41)	0.0008	3	(30)	0.4 (0.09, 1.46)	0.14	0	0	NA	0.13
Chills
No (n = 230)	44	19	1.0		124	(54)	1.0		24	10	1.0
Yes (n = 60)	19	32	2.0 (1.04, 3.70)	0.04	29	(48)	0.8 (0.45, 1.41)	0.44	7	12	1.1 (0.46, 2.77)	0.79
Hospitalization
No (n = 147)	27	(18)	1.0		78	(53)	1.0		17	(12)	1.0
Yes (n = 147)	37	(25)	1.5 (0.85, 2.62)	.16	77	(52)	1.0 (0.62, 1.54)	.91	14	(10)	0.8 (0.38, 1.70)	.57
Age (years)
0-18 (n = 148)	37	(25)	1.0		76	(51)	1.0		14	(9)	1.0
19-64 (n = 172)	32	(19)	0.7 (0.41, 1.17)	.16	93	(54)	1.1 (0.72, 1.73)	.63	20	(12)	1.3 (0.61, 2.59)	.53
Gender
Female (n = 171)	40	(23)	1.0		78	(46)	1.0		24	(14)	1.0
Male (n = 149)	29	(19)	0.8 (0.46, 1.36)	.39	91	(61)	1.9 (1.20, 2.92)	.006	10	(7)	0.4 (0.20, 0.95)	.03

Differences in the distribution of clades as measured by clinical data and bacterial characteristics were tested using the likelihood ratio c² (1 degree of freedom); odds ratios (OR), 95% confidence intervals (95% CI), and P values (P) were obtained based on these distributions.

*Percentages and associations are relative to all other clades combined; clade 9 strains were omitted from the analysis. Only one strain per outbreak or cluster was used in the analyses.

†Number varies depending on characteristic because some data were missing.

Table 4. Logistic regression results identifying predictors of hemolytic uremic syndrome (HUS) and infection with various Escherichia coli O157 clades among 333 Michigan patients

	HUS*		Clade 8†		Clade 2†		Clade 7†
Predictors	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Bloody diarrhea	0.8 (0.08, 8.50)	0.88	1.5 (0.63, 3.51)	0.36	1.6 (0.40, 1.14)	0.15	0.1 (0.06, 0.35)	<0.0001
Abdominal pain	0.4 (0.07, 1.85)	0.22	2.0 (0.79, 5.07)	0.14	1.2 (0.62, 2.23)	0.61	0.5 (0.19, 1.28)	0.15
HUS	--	--	7.0 (1.58, 31.31)	0.01	0.5 (0.11, 1.92)	0.29	NA	0.13
Chills	2.6 (0.37, 19.07)	0.33	2.0 (0.94, 4.32)	0.07	0.7 (0.38, 1.40)	0.34	1.6 (0.55, 4.77)	0.39
Hospitalization	4.7 (0.79, 27.65)	0.09	1.5 (0.79, 2.74)	0.23	0.9 (0.55, 1.49)	0.70	1.1 (0.48, 2.64)	0.78
Age (0-18 years)	16.70 (1.61, 172.78)	0.02	2.0 (1.04, 3.82)	0.04	0.7 (0.40, 1.14)	0.15	1.0 (0.42, 2.34)	0.97
Female	1.1 (0.25, 4.60)	0.93	1.2 (0.64, 2.16)	0.60	0.6 (0.34, 0.92)	0.02	1.9 (0.77, 4.44)	0.17
Clade 8 infection	6.1 (1.25, 29.94)	0.03	--	--	--	--	--	--
Clade 2 infection	0.5 (0.11, 2.32)	0.38	--	--	--	--	--	--

*Models used those without HUS as the reference group and were adjusted for bloody diarrhea, abdominal pain, diarrhea, chills, body aches, hospitalization, age, and gender.

†Models used those infected with all other clades except clade 9 as the reference group and were adjusted for bloody diarrhea, abdominal pain, diarrhea, chills, body aches, HUS, hospitalization, age, and gender.

This Article Abstract Full Text Services Email this article to a colleague Alert me to new issues of the journal Request Copyright Permission Citing Articles Citing Articles via CrossRef

Current Issue | Archives | Online Submission | Info for Authors | Editorial Board | About Subscribe | Advertise | Contact | Site Map Copyright © 2008 by the National Academy of Sciences