Phylogenetic relationship and virulence composition of Escherichia coli O26:H11 cattle and human strain collections in Scotland; 2002-2020

O26 is the commonest non-O157 Shiga toxin (stx)-producing Escherichia coli serogroup reported in human infections worldwide. Ruminants, particularly cattle, are the primary reservoir source for human infection. In this study, we compared the whole genomes and virulence profiles of O26:H11 strains (n = 99) isolated from Scottish cattle with strains from human infections (n = 96) held by the Scottish Escherichia coli O157/STEC Reference Laboratory, isolated between 2002 and 2020. Bovine strains were from two national cross-sectional cattle surveys conducted between 2002-2004 and 2014-2015. A maximum likelihood phylogeny was constructed from a core-genome alignment with the O26:H11 strain 11368 reference genome. Genomes were screened against a panel of 2,710 virulence genes using the Virulence Finder Database. All stx-positive bovine O26:H11 strains belonged to the ST21 lineage and were grouped into three main clades. Bovine and human source strains were interspersed, and the stx subtype was relatively clade-specific. Highly pathogenic stx2a-only ST21 strains were identified in two herds sampled in the second cattle survey and in human clinical infections from 2010 onwards. The closest pairwise distance was 9 single-nucleotide polymorphisms (SNPs) between Scottish bovine and human strains and 69 SNPs between the two cattle surveys. Bovine O26:H11 was compared to public EnteroBase ST29 complex genomes and found to have the greatest commonality with O26:H11 strains from the rest of the UK, followed by France, Italy, and Belgium. Virulence profiles of stx-positive bovine and human strains were similar but more conserved for the stx2a subtype. O26:H11 stx-negative ST29 (n = 17) and ST396 strains (n = 5) were isolated from 19 cattle herds; all were eae-positive, and 10 of these herds yielded strains positive for ehxA, espK, and Z2098, gene markers suggestive of enterohaemorrhagic potential. There was a significant association (p < 0.001) between nucleotide sequence percent identity and stx status for the bacteriophage insertion site genes yecE for stx2 and yehV for stx1. Acquired antimicrobial resistance genes were identified in silico in 12.1% of bovine and 17.7% of human O26:H11 strains, with sul2, tet, aph(3″), and aph(6″) being most common. This study describes the diversity among Scottish bovine O26:H11 strains and investigates their relationship to human STEC infections.

Prevalence of Shiga Toxin-Producing Escherichia coli O157 in Wild Scottish Deer with High Human Pathogenic Potential

Shiga toxin-producing E. coli (STEC) infections associated with wildlife are increasing globally, highlighting many 'spillover' species as important reservoirs for these zoonotic pathogens. A human outbreak of STEC serogroup O157 in 2015 in Scotland, associated with the consumption of venison meat products, highlighted several knowledge gaps, including the prevalence of STEC O157 in Scottish wild deer and the potential risk to humans from wild deer isolates. In this study, we undertook a nationwide survey of wild deer in Scotland and determined that the prevalence of STEC O157 in wild deer is low 0.28% (95% confidence interval = 0.06-0.80). Despite the low prevalence of STEC O157 in Scottish wild deer, identified isolates were present in deer faeces at high levels (>104 colony forming units/g faeces) and had high human pathogenic potential based on whole genome sequencing and virulence gene profiling. A retrospective epidemiological investigation also identified one wild deer isolate from this study as a possible source of a Scottish human outbreak in 2017. These results emphasise the importance of food hygiene practices during the processing of wild deer carcasses for human consumption.

Analysis of Escherichia coli O157 strains in cattle and humans between Scotland and England & Wales: implications for human health

For the last two decades, the human infection frequency of Escherichia coli O157 (O157) in Scotland has been 2.5-fold higher than in England and Wales. Results from national cattle surveys conducted in Scotland and England and Wales in 2014/2015 were combined with data on reported human clinical cases from the same time frame to determine if strain differences in national populations of O157 in cattle could be associated with higher human infection rates in Scotland. Shiga toxin subtype (Stx) and phage type (PT) were examined within and between host (cattle vs human) and nation (Scotland vs England and Wales). For a subset of the strains, whole genome sequencing (WGS) provided further insights into geographical and host association. All three major O157 lineages (I, II, I/II) and most sub-lineages (Ia, Ib, Ic, IIa, IIb, IIc) were represented in cattle and humans in both nations. While the relative contribution of different reservoir hosts to human infection is unknown, WGS analysis indicated that the majority of O157 diversity in human cases was captured by isolates from cattle. Despite comparable cattle O157 prevalence between nations, strain types were localized. PT21/28 (sub-lineage Ic, Stx2a+) was significantly more prevalent in Scottish cattle [odds ratio (OR) 8.7 (2.3-33.7; P<0.001] and humans [OR 2.2 (1.5-3.2); P<0.001]. In England and Wales, cattle had a significantly higher association with sub-lineage IIa strains [PT54, Stx2c; OR 5.6 (1.27-33.3); P=0.011] while humans were significantly more closely associated with sub-lineage IIb [PT8, Stx1 and Stx2c; OR 29 (4.9-1161); P<0.001]. Therefore, cattle farms in Scotland were more likely to harbour Stx2a+O157 strains compared to farms in E and W (P<0.001). There was evidence of limited cattle strain migration between nations and clinical isolates from one nation were more similar to cattle isolates from the same nation, with sub-lineage Ic (mainly PT21/28) exhibiting clear national association and evidence of local transmission in Scotland. While we propose the higher rate of O157 clinical cases in Scotland, compared to England and Wales, is a consequence of the nationally higher level of Stx2a+O157 strains in Scottish cattle, we discuss the multiple additional factors that may also contribute to the different infection rates between these nations.

Bacteriological Survey of Fresh Minced Beef on Sale at Retail Outlets in Scotland in 2019: Three Foodborne Pathogens, Hygiene Process Indicators, and Phenotypic Antimicrobial Resistance

ABSTRACT

The health and economic burden of foodborne illness is high, with approximately 2.4 million cases occurring annually in the United Kingdom. A survey to understand the baseline microbial quality and prevalence of food-related hazards of fresh beef mince on retail sale could inform risk assessment, management, and communication to ensure the safety of this commodity. In such a survey, a two-stage sampling design was used to reflect variations in population density and the market share of five categories of retail outlets in Scotland. From January to December 2019, 1,009 fresh minced beef samples were collected from 15 geographic areas. The microbial quality of each sample was assessed using aerobic colony count and Escherichia coli count. Samples were cultured for Campylobacter and Salmonella, and PCR was used to detect target genes (stx1 all variants, stx2 a to g, and rfbO157) for Shiga toxin-producing E. coli (STEC). The presence of viable E. coli O157 and STEC in samples with a positive PCR signal was confirmed via culture and isolation. Phenotypic antimicrobial sensitivity patterns of cultured pathogens and 100 E. coli isolates were determined, mostly via disk diffusion. The median aerobic colony count and E. coli counts were 6.4 × 105 (interquartile range, 6.9 × 104 to 9.6 × 106) and <10 CFU/g (interquartile range, <10 to 10) of minced beef, respectively. The prevalence was 0.1% (95% confidence interval [CI], 0 to 0.7%) for Campylobacter, 0.3% (95% CI, 0 to 1%) for Salmonella, 22% (95% CI, 20 to 25%) for PCR-positive STEC, and 4% (95% CI, 2 to 5%) for culture-positive STEC. The evidence for phenotypic antimicrobial resistance detected did not give cause for concern, mainly occurring in a few E. coli isolates as single nonsusceptibilities to first-line active substances. The low prevalence of pathogens and phenotypic antimicrobial resistance is encouraging, but ongoing consumer food safety education is necessary to mitigate the residual public health risk.

HIGHLIGHTS

Shiga toxin-producing Escherichia coli clonal complex 32, including serotype O145:H28, in the UK and Ireland

Introduction. Shiga toxin-producing Escherichia coli (STEC) O157:H7 has been the most clinically significant STEC serotype in the UK for over four decades. Over the last 10 years we have observed a decrease in STEC O157:H7 and an increase in non-O157 STEC serotypes, such as O145:H28.Gap Statement. Little is known about the microbiology and epidemiology of STEC belonging to CC32 (including O145:H28) in the UK. The aim of this study was to integrate genomic data with patient information to gain a better understanding of the virulence, disease severity, epidemic risk assessment and population structure of this clinically significant clonal complex.Methodology. Isolates of E. coli belonging to CC32 (n=309) in the archives of public health agencies in the UK and Ireland were whole-genome-sequenced, virulence-profiled and integrated with enhanced surveillance questionnaire (ESQ) data, including exposures and disease severity.Results. Overall, diagnoses of STEC belonging to CC32 (290/309, 94 %) in the UK have increased every year since 2014. Most cases were female (61 %), and the highest proportion of cases belonged to the 0-4 age group (53/211,25 %). The frequency of symptoms of diarrhoea (92 %), abdominal pain (84 %), blood in stool (71 %) and nausea (51 %) was similar to that reported in cases of STEC O157:H7, although cases of STEC CC32 were more frequently admitted to hospital (STEC CC32 48 % vs O157:H7  34 %) and/or developed haemolytic uraemic syndrome (HUS) (STEC CC32 9 % vs O157:H7 4 %).The majority of STEC isolates (268/290, 92 %) had the stx2a/eae virulence gene combination, most commonly associated with progression to STEC HUS. There was evidence of person-to-person transmission and small, temporally related, geographically dispersed outbreaks, characteristic of foodborne outbreaks linked to nationally distributed products.Conclusion. We recommend more widespread use of polymerase chain reaction (PCR) for the detection of all STEC serogroups, the development of consistent strategies for the follow-up testing of PCR-positive faecal specimens, the implementation of more comprehensive and standardized collection of epidemiological data, and routine sharing of sequencing data between public health agencies worldwide.

Epidemiology and genomic analysis of Shiga toxin-producing Escherichia coli clonal complex 165 in the UK

Introduction. Shiga toxin-producing Escherichia coli (STEC) is a zoonotic, foodborne gastrointestinal pathogen that has the potential to cause severe clinical outcomes, including haemolytic uraemic syndrome (HUS). STEC-HUS is the leading cause of renal failure in children and can be fatal. Over the last decade, STEC clonal complex 165 (CC165) has emerged as a cause of STEC-HUS.Gap Statement. There is a need to understand the pathogenicity and prevalence of this emerging STEC clonal complex in the UK, to facilitate early diagnosis, improve clinical management, and prevent and control outbreaks.Aim. The aim of this study was to characterize CC165 through identification of virulence factors (VFs) and antimicrobial resistance (AMR) determinants in the genome and to integrate the genome data with the available epidemiological data to better understand the incidence and pathogenicity of this clonal complex in the UK.Methodology. All isolates belonging to CC165 in the archives at the UK public health agencies were sequenced and serotyped, and the virulence gene and AMR profiles were derived from the genome using PHE bioinformatics pipelines and the Centre for Genomic Epidemiology virulence database.Results. There were 48 CC165 isolates, of which 43 were STEC, four were enteropathogenic E. coli (EPEC) and one E. coli. STEC serotypes were predominately O80:H2 (n=28), and other serotypes included O45:H2 (n=9), O55:H9 (n=4), O132:H2 (n=1) and O180:H2 (n=1). All but one STEC isolate had Shiga toxin (stx) subtype stx2a or stx2d and 47/48 isolates had the eae gene encoding intimin involved in the intimate attachment of the bacteria to the human gut mucosa. We detected extra-intestinal virulence genes including those associated with iron acquisition (iro) and serum resistance (iss), indicating that this pathogen has the potential to translocate to extra-intestinal sites. Unlike other STEC clonal complexes, a high proportion of isolates (93%, 40/43) were multidrug-resistant, including resistance to aminoglycosides, beta-lactams, chloramphenicol, sulphonamides, tetracyclines and trimethoprim.Conclusion. The clinical significance of this clonal complex should not be underestimated. Exhibiting high levels of AMR and a combination of STEC and extra-intestinal pathogenic E. coli (ExPEC) virulence profiles, this clonal complex is an emerging threat to public health.

Genome structural variation in Escherichia coli O157:H7

The human zoonotic pathogen Escherichia coli O157:H7 is defined by its extensive prophage repertoire including those that encode Shiga toxin, the factor responsible for inducing life-threatening pathology in humans. As well as introducing genes that can contribute to the virulence of a strain, prophage can enable the generation of large-chromosomal rearrangements (LCRs) by homologous recombination. This work examines the types and frequencies of LCRs across the major lineages of the O157:H7 serotype. We demonstrate that LCRs are a major source of genomic variation across all lineages of E. coli O157:H7 and by using both optical mapping and Oxford Nanopore long-read sequencing prove that LCRs are generated in laboratory cultures started from a single colony and that these variants can be recovered from colonized cattle. LCRs are biased towards the terminus region of the genome and are bounded by specific prophages that share large regions of sequence homology associated with the recombinational activity. RNA transcriptional profiling and phenotyping of specific structural variants indicated that important virulence phenotypes such as Shiga-toxin production, type-3 secretion and motility can be affected by LCRs. In summary, E. coli O157:H7 has acquired multiple prophage regions over time that act to continually produce structural variants of the genome. These findings raise important questions about the significance of this prophage-mediated genome contingency to enhance adaptability between environments.

Applying phylogenomics to understand the emergence of Shiga-toxin-producing Escherichia coli O157:H7 strains causing severe human disease in the UK

Shiga-toxin-producing Escherichia coli (STEC) O157:H7 is a recently emerged zoonotic pathogen with considerable morbidity. Since the emergence of this serotype in the 1980s, research has focussed on unravelling the evolutionary events from the E. coli O55:H7 ancestor to the contemporaneous globally dispersed strains observed today. In this study, the genomes of over 1000 isolates from both human clinical cases and cattle, spanning the history of STEC O157:H7 in the UK, were sequenced. Phylogenetic analysis revealed the ancestry, key acquisition events and global context of the strains. Dated phylogenies estimated the time to evolution of the most recent common ancestor of the current circulating global clone to be 175 years ago. This event was followed by rapid diversification. We show the acquisition of specific virulence determinates has occurred relatively recently and coincides with its recent detection in the human population. We used clinical outcome data from 493 cases of STEC O157:H7 to assess the relative risk of severe disease including haemolytic uraemic syndrome from each of the defined clades in the population and show the dramatic effect Shiga toxin repertoire has on virulence. We describe two strain replacement events that have occurred in the cattle population in the UK over the last 30 years, one resulting in a highly virulent strain that has accounted for the majority of clinical cases in the UK over the last decade. There is a need to understand the selection pressures maintaining Shiga-toxin-encoding bacteriophages in the ruminant reservoir and the study affirms the requirement for close surveillance of this pathogen in both ruminant and human populations.

E. coli O157 on Scottish cattle farms: evidence of local spread and persistence using repeat cross-sectional data

Background

Escherichia coli (E. coli) O157 is a virulent zoonotic strain of enterohaemorrhagic E. coli. In Scotland (1998-2008) the annual reported rate of human infection is 4.4 per 100,000 population which is consistently higher than other regions of the UK and abroad. Cattle are the primary reservoir. Thus understanding infection dynamics in cattle is paramount to reducing human infections.

A large database was created for farms sampled in two cross-sectional surveys carried out in Scotland (1998 - 2004). A statistical model was generated to identify risk factors for the presence of E. coli O157 on farms. Specific hypotheses were tested regarding the presence of E. coli O157 on local farms and the farms previous status. Pulsed-field gel electrophoresis (PFGE) profiles were further examined to ascertain whether local spread or persistence of strains could be inferred.

Results

The presence of an E. coli O157 positive local farm (average distance: 5.96km) in the Highlands, North East and South West, farm size and the number of cattle moved onto the farm 8 weeks prior to sampling were significant risk factors for the presence of E. coli O157 on farms. Previous status of a farm was not a significant predictor of current status (p = 0.398). Farms within the same sampling cluster were significantly more likely to be the same PFGE type (p < 0.001), implicating spread of strains between local farms. Isolates with identical PFGE types were observed to persist across the two surveys, including 3 that were identified on the same farm, suggesting an environmental reservoir. PFGE types that were persistent were more likely to have been observed in human clinical infections in Scotland (p < 0.001) from the same time frame.

Conclusions

The results of this study demonstrate the spread of E. coli O157 between local farms and highlight the potential link between persistent cattle strains and human clinical infections in Scotland. This novel insight into the epidemiology of Scottish E. coli O157 paves the way for future research into the mechanisms of transmission which should help with the design of control measures to reduce E. coli O157 from livestock-related sources.

Temporal and spatial patterns of bovine Escherichia coli O157 prevalence and comparison of temporal changes in the patterns of phage types associated with bovine shedding and human E. coli O157 cases in Scotland between 1998-2000 and 2002-2004

Background

Escherichia coli O157 is an important cause of acute diarrhoea, haemorrhagic colitis and, especially in children, haemolytic uraemic syndrome (HUS). Incidence rates for human E. coli O157 infection in Scotland are higher than most other United Kingdom, European and North American countries. Cattle are considered the main reservoir for E. coli O157. Significant associations between livestock related exposures and human infection have been identified in a number of studies.

Results

Animal Studies: There were no statistically significant differences (P = 0.831) in the mean farm-level prevalence between the two studies (SEERAD: 0.218 (95%CI: 0.141-0.32); IPRAVE: 0.205 (95%CI: 0.135-0.296)). However, the mean pat-level prevalence decreased from 0.089 (95%CI: 0.075-0.105) to 0.040 (95%CI: 0.028-0.053) between the SEERAD and IPRAVE studies respectively (P < 0.001). Highly significant (P < 0.001) reductions in mean pat-level prevalence were also observed in the spring, in the North East and Central Scotland, and in the shedding of phage type (PT) 21/28. Human Cases: Contrasting the same time periods, there was a decline in the overall comparative annual reported incidence of human cases as well as in all the major PT groups except 'Other' PTs. For both cattle and humans, the predominant phage type between 1998 and 2004 was PT21/28 comprising over 50% of the positive cattle isolates and reported human cases respectively. The proportion of PT32, however, was represented by few (<5%) of reported human cases despite comprising over 10% of cattle isolates. Across the two studies there were differences in the proportion of PTs 21/28, 32 and 'Other' PTs in both cattle isolates and reported human cases; however, only differences in the cattle isolates were statistically significant (P = 0.002).

Conclusion

There was no significant decrease in the mean farm-level prevalence of E. coli O157 between 1998 and 2004 in Scotland, despite significant declines in mean pat-level prevalence. Although there were declines in the number of human cases between the two study periods, there is no statistically significant evidence that the overall rate (per 100,000 population) of human E. coli O157 infections in Scotland over the last 10 years has altered. Comparable patterns in the distribution of PTs 21/28 and 32 between cattle and humans support a hypothesized link between the bovine reservoir and human infections. This emphasizes the need to apply and improve methods to reduce bovine shedding of E. coli O157 in Scotland where rates appear higher in both cattle and human populations, than in other countries.

Genetic diversity among clonal lineages within Escherichia coli O157:H7 stepwise evolutionary model

Molecular characterization and subtyping show genetic diversities within clonal complexes.

Escherichia coli O157:H7 variants were examined for trait mutations and by molecular subtyping to better define clonal complexes postulated on the O157:H7 evolution model. Strains of β-glucuronidase–positive, sorbitol-negative O157:H7 isolated in United States and Japan were identical to A5 clonal strain and shared sequence type (ST)–65 by multilocus sequence typing (MLST); thus, they belong in A5. However, these strains exhibited pulsed-field gel electrophoresis (PFGE) profile differences that suggested genomic divergence between populations. Sorbitol-fermenting O157 (SFO157) strains from Finland, Scotland, and Germany were identical to A4 clonal strain and belong in A4. Some SFO157 strains, isolated years apart and from different countries, had identical PFGE profiles, suggesting a common origin. Despite similarities, some Finnish and Scottish and all of the German strains have ST-75 (“German clone”), whereas others have ST-76, a new variant (“Scottish clone”). MLST of strains in other clonal complexes also discriminated strains thought to be identical and showed that genetic differences will further distinguish clonal populations into subclones.

Risk factors for the presence of high-level shedders of Escherichia coli O157 on Scottish farms

Escherichia coli O157 infections are the cause of sporadic or epidemic cases of often bloody diarrhea that can progress to hemolytic uremic syndrome (HUS), a systematic microvascular syndrome with predominantly renal and neurological complications. HUS is responsible for most deaths associated with E. coli O157 infection. From March 2002 to February 2004, approximately 13,000 fecal pat samples from 481 farms with finishing/store cattle throughout Scotland were examined for the presence of E. coli O157. A total of 441 fecal pats from 91 farms tested positive for E. coli O157. From the positive samples, a point estimate for high-level shedders was identified using mixture distribution analysis on counts of E. coli O157. Models were developed based on the confidence interval surrounding this point estimate (high-level shedder, greater than 10(3) or greater than 10(4) CFU g(-1) feces). The mean prevalence on high-level-shedding farms was higher than that on low-level-shedding farms. The presence of a high-level shedder on a farm was found to be associated with a high proportion of low-level shedding, consistent with the possibility of a higher level of transmission. Analysis of risk factors associated with the presence of a high-level shedder on a farm suggested the importance of the pathogen and individual host rather than the farm environment. The proportion of high-level shedders of phage 21/28 was higher than expected by chance. Management-related risk factors that were identified included the type of cattle (female breeding cattle) and cattle stress (movement and weaning), as opposed to environmental factors, such as water supply and feed.

Herd-level risk factors associated with the presence of Phage type 21/28 E. coli O157 on Scottish cattle farms

Background

E. coli O157 is a bacterial pathogen that is shed by cattle and can cause severe disease in humans. Phage type (PT) 21/28 is a subtype of E. coli O157 that is found across Scotland and is associated with particularly severe human morbidity.

Methods

A cross-sectional survey of Scottish cattle farms was conducted in the period Feb 2002-Feb 2004 to determine the prevalence of E. coli O157 in cattle herds. Data from 88 farms on which E. coli O157 was present were analysed using generalised linear mixed models to identify risk factors for the presence of PT 21/28 specifically.

Results

The analysis identified private water supply, and northerly farm location as risk factors for PT 21/28 presence. There was a significant association between the presence of PT 21/28 and an increased number of E. coli O157 positive pat samples from a farm, and PT 21/28 was significantly associated with larger E. coli O157 counts than non-PT 21/28 E. coli O157.

Conclusion

PT 21/28 has significant risk factors that distinguish it from other phage types of E. coli O157. This finding has implications for the control of E. coli O157 as a whole and suggests that control could be tailored to target the locally dominant PT.

VTEC infections and livestock-related exposures in Scotland, 2004

Scotland generally reports higher rates of infection with verotoxin-producing Escherichia coli O157 (VTEC O157) than other parts of the United Kingdom

Rectal carriage of enterohemorrhagic Escherichia coli O157 in slaughtered cattle

Escherichia coli O157:H7 is an important cause of diarrhea, hemorrhagic colitis, and potentially fatal human illness. Cattle are considered a primary reservoir of infection, and recent experimental evidence has indicated that the terminal rectum is the principal site of bacterial carriage. To test this finding in naturally colonized animals, intact rectum samples from 267 cattle in 24 separate lots were obtained immediately after slaughter, and fecal material and mucosal surfaces were cultured for E. coli O157 by direct and enrichment methods. Two locations, 1 and 15 cm proximal to the recto-anal junction, were tested. In total, 35 animals were positive for E. coli O157 at at least one of the sites and 232 animals were negative as determined by all tests. The frequency of isolation and the numbers of E. coli O157 cells were higher at the site closer to the recto-anal junction, confirming our previous experimental findings. We defined low- and high-level carriers as animals with E. coli O157 levels of <1 x 10(3) CFU g(-1) or <1 x 10(3) CFU ml(-1) and animals with E. coli O157 levels of > or =1 x 10(3) CFU g(-1) or > or =1 x 10(3) CFU ml(-1) in feces or tissues, respectively. High-level carriage was detected in 3.7% of the animals (95% confidence interval, 1.8 to 6.8%), and carriage on the mucosal surface of the terminal rectum was associated with high-level fecal excretion. In summary, our results support previous work demonstrating that the mucosal epithelium in the bovine terminal rectum is an important site for E. coli O157 carriage in cattle. The data also support the hypothesis that high-level fecal shedding (> or =1 x 10(3) CFU g of feces(-1)) of enterohemorrhagic E. coli O157 results from colonization of this site.

Genetic typing of shiga toxin 2 variants of Escherichia coli by PCR-restriction fragment length polymorphism analysis

Shiga toxins Stx1 and Stx2 play a prominent role in the pathogenesis of Shiga toxin-producing Escherichia coli (STEC) infections. Several variants of the stx(2) gene, encoding Stx2, have been described. In this study, we developed a PCR-restriction fragment length polymorphism system for typing stx(2) genes of STEC strains. The typing system discriminates eight described variants and allows the identification of new stx(2) variants and STEC isolates carrying multiple stx(2) genes. A phylogenetic tree, based on the nucleotide sequences of the toxin-encoding genes, demonstrates that stx(2) sequences with the same PvuII HaeIII HincII AccI type generally cluster together.

Evaluation of the Escherichia coli threonine deaminase gene as a selectable marker for plant transformation

The initial step in the synthesis of isoleucine (Ile) is the conversion of threonine to alpha-ketobutyrate. This reaction is carried out by threonine deaminase (TD), which is feedback-regulated by Ile. Mutations in TD that manifest insensitivity to Ile feedback inhibition result in intracellular accumulation of Ile. Previous reports have shown that in planta expression of the wild-type Escherichia coli TD, ilvA, or an Ile-insensitive mutant designated ilvA-466, increased cellular concentrations of Ile. A structural analog of Ile, l-O-methylthreonine (OMT), is able to compete effectively with Ile during translation and induce cell death. It has been postulated that OMT could therefore be utilized as an effective selective agent in plant engineering studies. To test this concept, we designed two binary plasmids that harbored an nptII cassette and either the wild-type ilvA or mutant ilvA-466. The ilvA coding sequences were fused to a plastid transit peptide down stream of a modified 35S CaMV promoter. Tobacco transformations were set up implementing a selection protocol based on either kanamycin or OMT. The ilvA gene was effectively utilized as a selectable marker gene to identify tobacco transformants when coupled with OMT as the selection agent. However, the transformation efficiency was substantially lower than that observed with nptII using kanamycin as the selection agent. Moreover, in a subset of the ilvA transformants and in a majority of the ilvA-466 transgenic lines, a severe off-type was observed under greenhouse conditions that correlated with increased levels of expression of the ilvA transgene.

Discovering patterns in Plasmodium falciparum genomic DNA

A method has been developed for discovering patterns in DNA sequences. Loosely based on the well-known Lempel Ziv model for text compression, the model detects repeated sequences in DNA. The repeats can be forward or inverted, and they need not be exact. The method is particularly useful for detecting distantly related sequences, and for finding patterns in sequences of biased nucleotide composition, where spurious patterns are often observed because the bias leads to coincidental nucleotide matches. We show here the utility of the method by applying it to genomic sequences of Plasmodium falciparum. A single scan of chromosomes 2 and 3 of P. falciparum, using our method and no other a priori information about the sequences, reveals regions of low complexity in both telomeric and central regions, long repeats in the subtelomeric regions, and shorter repeat areas in dense coding regions. Application of the method to a recently sequenced contig of chromosome 10 that has a particularly biased base composition detects a long internal repeat more readily than does the conventional dot matrix plot. Space requirements are linear, so the method can be used on large sequences. The observed repeat patterns may be related to large-scale chromosomal organization and control of gene expression. The method has general application in detecting patterns of potential interest in newly sequenced genomic material.

Fast, optimal alignment of three sequences using linear gap costs

Alignment algorithms can be used to infer a relationship between sequences when the true relationship is unknown. Simple alignment algorithms use a cost function that gives a fixed cost to each possible point mutation-mismatch, deletion, insertion. These algorithms tend to find optimal alignments that have many small gaps. It is more biologically plausible to have fewer longer gaps rather than many small gaps in an alignment. To address this issue, linear gap cost algorithms are in common use for aligning biological sequence data. More reliable inferences are obtained by aligning more than two sequences at a time. The obvious dynamic programming algorithm for optimally aligning k sequences of length n runs in O(n(k)) time. This is impractical if k>/=3 and n is of any reasonable length. Thus, for this problem there are many heuristics for aligning k sequences, however, they are not guaranteed to find an optimal alignment. In this paper, we present a new algorithm guaranteed to find the optimal alignment for three sequences using linear gap costs. This gives the same results as the dynamic programming algorithm for three sequences, but typically does so much more quickly. It is particularly fast when the (three-way) edit distance is small. Our algorithm uses a speed-up technique based on Ukkonen's greedy algorithm (Ukkonen, 1983) which he presented for two sequences and simple costs.

Clinical effectiveness: testicular self-examination

Evidence must be used to support a change in practice. Incorporating evidence into practice is essential for clinical effectiveness. The findings should be disseminated to others involved. The effect on patients should be evaluated.

Sequence complexity for biological sequence analysis

A new statistical model for DNA considers a sequence to be a mixture of regions with little structure and regions that are approximate repeats of other subsequences, i.e. instances of repeats do not need to match each other exactly. Both forward- and reverse-complementary repeats are allowed. The model has a small number of parameters which are fitted to the data. In general there are many explanations for a given sequence and how to compute the total probability of the data given the model is shown. Computer algorithms are described for these tasks. The model can be used to compute the information content of a sequence, either in total or base by base. This amounts to looking at sequences from a data-compression point of view and it is argued that this is a good way to tackle intelligent sequence analysis in general.

Genetic heterogeneity of Escherichia coli O157:H7 in Scotland and its utility in strain subtyping

From April 1994 to March 1995, seven outbreaks of Escherichia coli O157:H7 infection occurred throughout Scotland, including the largest milk-borne outbreak to date worldwide. Various vehicles of infection were identified, and there were 144 confirmed cases in total. All isolates associated with the outbreaks were subjected to detailed subtyping: phage typing, testing for carriage of verotoxin genes (VT), and pulsed-field gel electrophoresis. The outbreak strains were of three different phage types (2, 4, and 28). Those of phage type 2 and 28 were VT1-/VT2+, those of phage type 4 were VT1+/VT2+. To discriminate outbreak-associate isolates from the high sporadic background, real-time pulsed-field gel electrophoresis analyses were performed. The results demonstrated that, within each of the seven outbreak groups, the macrorestriction profiles observed were indistinguishable, whereas profiles for sporadic isolates were not. The consistent genetic heterogeneity observed within the Scottish Escherichia coli O157 population can be exploited in epidemiological investigations.

rbcL Transcript levels in tobacco plastids are independent of light: reduced dark transcription rate is compensated by increased mRNA stability

The plastid rbcL gene, encoding the large subunit of ribulose-1, 5-bisphosphate carboxylase, in higher plants is transcribed from a sigma70 promoter by the eubacterial-type RNA polymerase. To identify regulatory elements outside of the rbcL -10/-35 promoter core, we constructed transplastomic tobacco plants with uidA reporter genes expressed from rbcL promoter derivatives. Promoter activity was characterized by measuring steady state levels of uidA mRNA on RNA gel blots and by measuring promoter strength in run-on transcription assays. We report here that the rbcL core promoter is sufficient to obtain wild-type rates of transcription. Furthermore, the rates of transcription were up to 10-fold higher in light-grown leaves than in dark-adapted plants. Although the rates of transcription were lower in the dark, rbcL mRNA accumulated to similar levels in light-grown and dark-adapted leaves. Accumulation of uidA mRNA from most rbcL promoter deletion derivatives directly reflected the relative rates of transcription: high in the light-grown and low in the dark-adapted leaves. However, uidA mRNA accumulated to high levels in a light-independent fashion as long as a segment encoding a stem-loop structure in the 5' untranslated region was included in the promoter construct. This finding indicates that lower rates of rbcL transcription in the dark are compensated by increased mRNA stability.

Discovering simple DNA sequences by compression

An information-theoretic DNA compression scheme devised by Milosavljevic and Jurka (1993) has been used in many places in the literature for both the discovery of new genes and the compression of DNA. Their compression method applies an encoding of previously occurring runs. They use 5 different code-words: four being the DNA bases, A, C, G and T, and the other being a pointer to a previously occurring run. They advocate a code-word of length log2 5 for each of these and then encoding a run by a code-word of length 2 x log2 n, where n is the length of the sequence. This scheme encodes the start of the sequence with a code-word of length log2 n and likewise encodes the end of the sequence with a code-word of length log2 n. In this paper, we show the above coding scheme to be inefficient in various ways and improve upon it so that it can compress DNA. We discuss our implementation of various schemes some of which run in linear time.

An MML classification of protein structure that knows about angles and sequence

The MML classification program, Snob, deals with mixture modelling (or clustering) of circular data. It has recently been extended to do Markov modelling of the serial correlation between clusters such as modelling the fact that a Helix cluster favours being followed by another Helix cluster. Such a model is better known as a Hidden Markov Model. The search for the most appropriate secondary structure classification of protein data is of significant importance and was addressed by Hunter and States (1992) using the Bayesian classifier, AutoClass, on Cartesian co-ordinate data of protein residues. Dowe et al. (1996) improved upon this earlier work by using Snob to cluster dihedral angle data, with the advantage that 3 x 3 = 9 Cartesian co-ordinates can be represented by the 2 orientation-invariant angles, phi and psi. The Hidden Markov Model used here is shown to be a more appropriate way again of modelling protein data and results in the selection of a simpler class model with 17 structure classes. We report on this classification, including the class transition matrix, and relate it back to the amino-acid sequence and the simple Helix, Beta, Turn classification. We find 3 types of Helix, 2 types of Beta and many types of Turn. The msot numerous Turn class defines a continuous flexible structure that is negatively correlated to all the other classes.

Compression of strings with approximate repeats

We describe a model for strings of characters that is loosely based on the Lempel Ziv model with the addition that a repeated substring can be an approximate match to the original substring; this is close to the situation of DNA, for example. Typically there are many explanations for a given string under the model, some optimal and many suboptimal. Rather than commit to one optimal explanation, we sum the probabilities over all explanations under the model because this gives the probability of the data under the model. The model has a small number of parameters and these can be estimated from the given string by an expectation-maximization (EM) algorithm. Each iteration of the EM algorithm takes O(n2) time and a few iterations are typically sufficient. O(n2) complexity is impractical for strings of more than a few tens of thousands of characters and a faster approximation algorithm is also given. The model is further extended to include approximate reverse complementary repeats when analyzing DNA strings. Tests include the recovery of parameter estimates from known sources and applications to real DNA strings.

Circular clustering of protein dihedral angles by Minimum Message Length

Early work on proteins identified the existence of helices and extended sheets in protein secondary structures, a high-level classification which remains popular today. Using the Snob program for information-theoretic Minimum Message Length (MML) classification, we are able to take the protein dihedral angles as determined by X-ray crystallography, and cluster sets of dihedral angles into groups. Previous work by Hunter and States has applied a similar Bayesian classification method, AutoClass, to protein data with site position represented by 3 Cartesian co-ordinates for each of the alpha-Carbon, beta-Carbon and Nitrogen, totalling 9 co-ordinates. By using the von Mises circular distribution in the Snob program, we are instead able to represent local site properties by the two dihedral angles, phi and psi. Since each site can be modelled as having 2 degrees of freedom, this orientation-invariant dihedral angle representation of the data is more compact than that of nine highly-correlated Cartesian co-ordinates. Using the information-theoretic message length concepts discussed in the paper, such a more concise model is more likely to represent the underlying generating process from which the data came. We report on the results of our classification, plotting the classes in (phi, psi) space; and introducing a symmetric information-theoretic distance measure to build a minimum spanning tree between the classes. We also give a transition matrix between the classes and note the existence of three classes in the region phi approximately -1.09 rad and psi approximately -0.75 rad which are close on the spanning tree and have high inter-transition probabilities. This gives rise to a tight, abundant and self-perpetuating structure.

Relationship between head growth and neurodevelopmental outcome of Malaysian very low birthweight infants during the 1st year of life

A prospective study was carried out to (i) compare head growth patterns of 103 very low birthweight (VLBW, < 1500 g) Malaysian infants and 98 normal birthweight (NBW, 2500- < 4500 g) controls during the 1st year of life; and (ii) examine the relationship between neurodevelopmental outcome at 1 year of age and occipito-frontal head circumferences (OFC) at birth and at 1 year of age in VLBW babies. When compared with those of NBW infants at birth, mid-infancy and 1 year of age, the mean OFC ratios (observed/expected OFC at 50th percentile) of VLBW infants were significantly lower (p < 0.001). Small-for-gestational-age (SGA) VLBW babies had significantly lower mean OFC ratios than their appropriate-for-gestational-age (AGA) VLBW counterparts at birth (p < 0.001), but this difference was no longer seen at mid-infancy or at 1 year of age. Logistic regression analysis showed that abnormal late neonatal cranial ultrasound findings (odds ratio 8.5, 95% confidence interval 4.12-22.07; p < 0.001) and each additional day of oxygen therapy (odds ratio 1.15, 95% confidence interval 1.00-4.45; p = 0.045) were significant risk factors associated with neurodevelopmental disability at 1 year of age, while mean OFC ratios at birth or at 1 year of age were not. Poor postnatal head growth per se did not predict disability, but probably reflected the consequences of "brain injury" as evidenced by abnormal brain scans.

Neurodevelopmental outcome of Malaysian very low birth weight infants: predictive value of cranial ultrasound appearances

The aim of the study was to determine the predictive value of cranial ultrasound scans done in the neonatal period for neurodevelopmental outcome of the Malaysian very low birthweight (VLBW, < 1500 grams) infants assessed at 12 months of corrected age. Of the 101 infants studied, 68 (67.3%) were neurodevelopmentally normal at one year of age, 18 (17.8%) had major and 15 (14.9%) had minor neurodevelopmental impairment. Neurodevelopmental outcome was normal in 66/88 (75.0%) infants who did not have severe intraventricular haemorrhage (IVH) or periventricular intraparenchymal echo densities (PVE) in the first week of life, and in 57/73 (78.1%) with uncomplicated scans at discharge. In contrast, 11/13 (84.6%) with parenchymal echo densities or severe intraventricular bleed in the early neonatal period and 17/28 (60.7%) with complicated scans at discharge had adverse sequelae. There was a significant association between lesions seen on cranial ultrasound in the neonatal period and subsequent neurodevelopmental impairment. Late neonatal ultrasound scans appear to be a better predictor of short-term neurodevelopmental outcome than early scans.

Comparison of morbidities in very low birthweight and normal birthweight infants during the first year of life in a developing country

OBJECTIVE

To compare the morbidities in the very low birthweight (VLBW; < 1500 g) and normal birthweight (NBW; > or = 2500 g) Malaysian infants during the first year of life.

METHODOLOGY

Prospective observational cohort study of consecutive surviving VLBW infants and randomly sampled NBW infants born in the Kuala Lumpur Maternity Hospital between 1 December 1989 and 31 December 1992. Infants were followed up regularly during the first year of life, after correction for prematurity.

RESULTS

Compared with NBW infants (n = 106), VLBW infants (n = 127) had significantly higher risk of failure to thrive (odds ratio [OR] = 8.0, 95% confidence intervals [CI]: 1.1 to 354.3), wheezing (OR = 3.7, 95% CI: 1.6 to 9.3), rehospitalization (OR = 2.3, 95% CI: 1.1 to 5.0), cerebral palsy (OR = 8.6, 95% CI: 2.0 to 77.6), neurosensory hearing loss (OR = 12.0, 95% CI: 1.7 to 513.6) and visual loss (7.9 vs 0%, P = 0.002). The mean mental developmental index (MDI) and mean psychomotor developmental index (PDI) at 1 year of age were significantly lower among VLBW infants (MDI 99 [SD = 28], PDI 89 [SD = 25]) than NBW infants (MDI 106 [SD = 18], PDI 101 [SD = 18]) (95% CI for difference between means being MDI: -14.1 to -1.7; and PDI: -17.6 to -6.0). Logistic regression analysis showed that among VLBW infants: (i) male sex, Malay ethnicity and bronchopulmonary dysplasia were significant risk factors associated with wheezing; (ii) longer duration of oxygen therapy during the neonatal period, seizures after the post-neonatal period and wheezing were significant risk factors associated with rehospitalization; and (iii) longer duration of oxygen therapy during the neonatal period was a significant risk factor associated with adverse neurodevelopmental outcome during the first year of life.

CONCLUSIONS

Compared with NBW infants, VLBW Malaysian infants had significantly higher risks of physical and neuro-developmental morbidities.

Long-term efficacy of continuing hepatitis B vaccination in infancy in two Gambian villages

In 1984, all non-immune children under the age of 5 years in the Gambian villages of Keneba and Manduar were vaccinated against hepatitis B virus (HBV). All children born in these villages since 1984 have been vaccinated in infancy. Despite a rapid fall in antibody concentrations, vaccine efficacy against HBV infection and chronic carriage of HBsAg has increased with time. Overall, vaccine efficacies in 1993 against HBV infection and chronic HBsAg carriage were 94.7% (95% Cl 93.0-96.0) and 95.3% (91.0-97.5), respectively. Breakthrough infections in vaccinated children largely originate from chronic HBsAg carriers. Thus, we tested 261 chronic carriers for HBV DNA and e antigen. The prevalence of these markers of infectivity, and the amount of HBV DNA, decreased greatly with age. Detailed studies of breakthrough infections over two 4-year periods revealed that in the second period there were fewer than half the expected numbers of infections. Our findings suggest that in Keneba and Manduar long-term vaccination is progressively decreasing HBV transmission by chronic carriers, since their infectivity diminishes with time.

The posterior probability distribution of alignments and its application to parameter estimation of evolutionary trees and to optimization of multiple alignments

How to sample alignments from their posterior probability distribution given two strings is shown. This is extended to sampling alignments of more than two strings. The result is first applied to the estimation of the edges of a given evolutionary tree over several strings. Second, when used in conjunction with simulated annealing, it gives a stochastic search method for an optimal multiple alignment.

Breakthrough infections and identification of a viral variant in Gambian children immunized with hepatitis B vaccine

Hepatitis B (HB) breakthrough infections, identified by the presence of HB core (c) antibody, were found in 32 of 358 Gambian children vaccinated with plasma-derived HB vaccine. Over 2 years, 15 of these children lost their HBc antibodies. These children had significantly higher HB surface antibody levels before infection than those who retained HBc antibodies. One child, who responded well to the vaccine, had HB viral DNA detected in the presence of HBs antibodies. The S gene sequence of this DNA showed nucleotide changes that resulted in an amino acid substitution at residue 141 (lysine to glutamic acid) of the surface antigen. This finding suggests the child was infected with a variant virus that was not neutralized by antibodies resulting from HB vaccination.

A fast algorithm for the optimal alignment of three strings

Ukkonen's (pair-wise) string alignment technique is extended to the problem of finding an optimal alignment for three strings. The resulting algorithm has worst-case time-complexity O(nd2) and space-complexity O(d3), where the string lengths are ñ and d is the three-way edit-distance based on tree-costs. In practice, the algorithm usually runs in O(n + d3) time. The algorithm is particularly fast when the strings are similar, in which case, d << n. Three-way alignment is an important special case in string alignment. Each internal node in an unrooted, binary evolutionary-tree has three neighbours. The algorithm presented can be used as an iterative step in a heuristic multiple-alignment program for more than three strings.

Normalization of affine gap costs used in optimal sequence alignment

It is shown how to normalize the costs of an alignment algorithm that employs affine or linear gap costs. The normalized costs are interpreted as the -log probabilities of the instructions of a finite-state edit-machine. This gives an explicit model relating sequences that can be linked to processes of mutation and evolution.

Reconstruction of strings past

A major use of string-alignment algorithms is to compare macromolecules that are thought to have evolved from a common ancestor to estimate the duration of, or the amount of mutation in, their separate evolution and to infer as much as possible about their most recent common ancestor. Minimum message length encoding, a method of inductive inference, is applied to the string-alignment problem. It leads to an alignment method that averages over all alignments in a weighted fashion. Experiments indicates that this method can recover the actual parameters of evolution with high accuracy and over a wide range of values, whereas the use of a single optimal alignment gives biased results.

Finite-state models in the alignment of macromolecules

Minimum message length encoding is a technique of inductive inference with theoretical and practical advantages. It allows the posterior odds-ratio of two theories or hypotheses to be calculated. Here it is applied to problems of aligning or relating two strings, in particular two biological macromolecules. We compare the r-theory, that the strings are related, with the null-theory, that they are not related. If they are related, the probabilities of the various alignments can be calculated. This is done for one-, three-, and five-state models of relation or mutation. These correspond to linear and piecewise linear cost functions on runs of insertions and deletions. We describe how to estimate parameters of a model. The validity of a model is itself an hypothesis and can be objectively tested. This is done on real DNA strings and on artificial data. The tests on artificial data indicate limits on what can be inferred in various situations. The tests on real DNA support either the three- or five-state models over the one-state model. Finally, a fast, approximate minimum message length string comparison algorithm is described.

Restriction site mapping for three or more enzymes

Restriction site mapping requires a generator to put forward possible maps and a constraint checker to reject false maps. Ideally these combine to give an algorithm which calculates a sound and complete solution set. Three algorithms for generation are presented and compared. Two decompose a multi-enzyme problem (greater than or equal to 3) into subproblems. The constraint checker is based on separation theory. Some insights into the extent of constraint checking involved in and feasibility of more checking for three or more enzymes are discussed. The trade-off between computation time and the soundness of the solution set is examined.

Minimum message length encoding and the comparison of macromolecules

A comparison of inductive inference known as minimum message length encoding is applied to string comparison in molecular biology. The question of whether or not two strings are related and, if so, of how they are related and the problem of finding a good theory of string mutation are treated as inductive inference problems. The method allows the posterior odds-ratio of two string alignments or of two models of string mutation to be computed. The connection between models of mutation and existing string alignment algorithms is made explicit. A fast minimum message length alignment algorithm is also described.