Statistical analysis

To understand the relationship between Campylobacter spp. and different sampling lo-

cations, and over time, a linear model was used to predict the influence of space and time. Simple linear regression analysis was carried out where, the total number of sam- ples collected per day, per site was 12 throughout the entire study period for both ducks and starlings. Site and time of sampling were used as explanatory variables. The infer- ences on the significance of space and time were drawn based on the probability values

obtained from the model. Means ofCampylobacterprevalence in ducks and starlings and

the means ofC. jejuniprevalence in ducks and starlings were compared and tested with

the function of analysis of variance (ANOVA).

Population diversity : Rarefaction

Rarefaction was performed by using the frequency of STs (as described by Gormley et al.

2008) found in ducks and starlings. The frequency distribution of eachC. jejuniST was

summarised at the sampling sites, species and seasonal level in an effort to describe how

theC. jejunipopulation varied according to sampling sites, species and time of the year.

The analysis was carried out using the contributed R package Vegan (R: A Language and

Environment for Statistical Computing, R Development Core Team, R Foundation for Sta- tistical Computing, Vienna, Austria, 2010, ISBN 3-900051-07-0, http://www.R-project.org

n.d.) that contains a rarefaction function. In rarefaction analysis, the horizontal axis of the plot represents the number of samples used for analysis and the vertical represents the diversity or the number of sequence types identified in the specified number of sam- ples. Diversity indices such as Simpson index 1-D and Shannon index were measured (as described in the PAST, PAleontological STatistics reference manual v2.13 by Hammer

2009-2011 [page number 40]) to analyse theC. jejuni population in ducks and starlings.

Simpson index 1-D measures the evenness of the community that has a scale from 0 to 1, where 0 indicates that all taxa are equally present and 1 indicates that one taxon dominates the community completely. Whereas, Shannon index takes the number of individuals as well as number of taxa into account that varies from value 0 for communities with only a single taxon to accounting for high values or frequency and values above 0 for communi- ties with many taxa, each with few individuals.

Population differentiation:

Analysis of molecular variance and Fst

Analysis of molecular variance (AMOVA) was performed to compare the effect of sam- pling site, host species and sampling period on the population differentiation and pop-

ulation structure of C. jejuni Arlequin v3.11 (Excoffier et al. 2005). In AMOVA, the

genetic structure of a given population is analysed by an analysis of variance framework. AMOVA analyses the variance of allelic frequencies within and between populations or groups. Analysis at different levels, referred to as ‘hierarchical analysis’ is carried out in AMOVA which divides the total variance into different covariance components such as within population, within groups among populations and inter-population differences (Excoffier et al. 2005). While analysing the genetic structure for different hierarchical levels, three hierarchical F-statistics are derived known as the fixation indices (expressed as components of AMOVA). The fixation indices include: Fst, a fixation index that mea- sures the variance among subpopulations relative to the total variance, Fsc, that measures

3.2 Experimental procedures 61

variance among subpopulations within groups and Fct, the variance among groups rel- ative to the total variance. Fst quantifies the genetic differentiation among populations under comparison using an index ranging from 0 to 1. A zero value implies that there is no differentiation between populations and a value of one implies the two populations are completely separate (Wright 1965, 1978, 1984). Wright (1978) has further suggested the qualitative guidelines for the interpretation of Fst such as: (1)the range 0.0 to 0.05 may be considered as indicating small or limited genetic differentiation; (2) the range 0.05 to 0.15 indicates moderate genetic differentiation (3) the range 0.15 to 0.25 indicates great genetic differentiation and (4) the values of Fst above 0.25 indicate very great genetic dif-

ferentiation. The genetic distance ofC. jejunipopulations found in ducks and starlings at

different time periods and at different sampling sites were calculated using MLST allelic profiles. In turn, all these fixation indices facilitate inference on the gene flow between populations compared.

Minimum spanning tree Phylogenetic relationships among the STs were analysed using

Bionumerics v6.1.4 _{A minimum spanning tree (MST) was constructed using the allelic}

profile data set for ducks and starlings collected from March 2008 (the trial period in- cluded) to July 2009. MST is an alternative approach to show the relationships among isolates from bacterial populations (Prim 1957). A MST links the STs that are closely related within their lineages or CCs, identifying the most likely extant ancestral sequence type for the STs analysed. The most likely extant ancestral sequence types are known as the ‘consensus’ clones or the founder STs from which the clonal variants (STs) have descended (Feil & Chan 2001). The radial spread of STs from the consensus clones are reflected by a series of circles where the size of the circles represent the number of isolates

per ST (Prim 1957). The construction of a MST involves three main steps:5

1. The sub-division of the data into ’clonal complexes’.

2. The identification of ancestral sequence types. Inferring likely patterns of descent within each clonal complex.

Clonal complexes are multilocus sequence types in which every sequence type shares at least 5 loci in common with at least one other member of the group. Ancestral se- quence types (or ’consensus clones’ or founder strains) are identified based on the seven

4_{Applied Maths URL:http://applied-maths.com/bionumerics/bionumerics.htm}

loci where an ancestral sequence type differs from the highest number of other sequence types in the clonal complex at only one locus out of seven that defines the highest num- ber of single-locus variants, or SLVs. Single-locus variants are identical to the ancestral sequence type at 6 loci, but differ at the seventh while the double locus variants (DLVs) vary at two loci. Once the ancestral sequence types are assigned, strains are assigned ac- cording to their relationships with their respective ancestral sequence types of the clonal complex, where all the SLVs and DLVs are associated with their corresponding ancestral sequence types. Typically, strains that share at least 5 loci in common are included in a clonal complex.