Abstract Levels of a predictor variable that are absent when a classification tree is grown can not be subject to an explicit splitting rule. This is an issue if these absent levels are present in a new observation for prediction. To date, there remains no satisfactory solution for absent levels in random forest models. Unlike missing data, absent levels are fully observed and known. Ordinal encoding of predictors allows absent levels to be integrated and used for prediction. Using a case study on source attribution of Campylobacter species using whole genome sequencing (WGS) data as predictors, we examine how target-agnostic versus target-based encoding of predictor variables with absent levels affects the accuracy of random forest models. We show that a target-based encoding approach using class probabilities, with absent levels designated the highest rank, is systematically biased, and that this bias is resolved by encoding absent levels according to the a priori hypothesis of equal class probability. We present a novel method of ordinal encoding predictors via principal coordinates analysis (PCO) which capitalizes on the similarity between pairs of predictor levels. Absent levels are encoded according to their similarity to each of the other levels in the training data. We show that the PCO-encoding method performs at least as well as the target-based approach and is not biased.

Disentangling the impact of the weather on transmission of infectious diseases is crucial for health protection, preparedness and prevention. Because weather factors are co-incidental and partly correlated, we have used geography to separate out the impact of individual weather parameters on other seasonal variables using campylobacteriosis as a case study. Campylobacter infections are found worldwide and are the most common bacterial food-borne disease in developed countries, where they exhibit consistent but country specific seasonality. We developed a novel conditional incidence method, based on classical stratification, exploiting the long term, high-resolution, linkage of approximately one-million campylobacteriosis cases over 20 years in England and Wales with local meteorological datasets from diagnostic laboratory locations. The predicted incidence of campylobacteriosis increased by 1 case per million people for every 5° (Celsius) increase in temperature within the range of 8°-15°. Limited association was observed outside that range. There were strong associations with day-length. Cases tended to increase with relative humidity in the region of 75-80%, while the associations with rainfall and wind-speed were weaker. The approach is able to examine multiple factors and model how complex trends arise, e.g. the consistent steep increase in campylobacteriosis in England and Wales in May-June and its spatial variability. This transparent and straightforward approach leads to accurate predictions without relying on regression models and/or postulating specific parameterisations. A key output of the analysis is a thoroughly phenomenological description of the incidence of the disease conditional on specific local weather factors. The study can be crucially important to infer the elusive mechanism of transmission of campylobacteriosis; for instance, by simulating the conditional incidence for a postulated mechanism and compare it with the phenomenological patterns as benchmark. The findings challenge the assumption, commonly made in statistical models, that the transformed mean rate of infection for diseases like campylobacteriosis is a mere additive and combination of the environmental variables.

Motivation: Large-scale whole-genome sequencing dataset-based studies are becoming increasingly common in pathogen surveillance and outbreak investigations. A highly discriminative and time-efficient bioinformatics tool is needed to transform large amounts of sequencing data into usable biological information. To replace the intuitive, yet inefficient, way of gene-by-gene allele calling algorithm, a new algorithm using genome-by-genome approach was developed. Results: Tests showed that the program equipped with the new algorithm achieved significant improvements in allele calling efficiency compared to a conventional gene-by-gene approach. The new program, Fast-GeP, rendered a fast and easy way to infer high-resolution genealogical relationships between bacterial isolates using whole-genome sequencing data. Availability and implementation: FAST-GeP is freely available from: https://github.com/jizhang-nz/fast-GeP. Supplementary information: Supplementary data are available at Bioinformatics online.

Zoonotic diseases are a major cause of morbidity, and productivity losses in both human and animal populations. Identifying the source of food-borne zoonoses (e.g. an animal reservoir or food product) is crucial for the identification and prioritisation of food safety interventions. For many zoonotic diseases it is difficult to attribute human cases to sources of infection because there is little epidemiological information on the cases. However, microbial strain typing allows zoonotic pathogens to be categorised, and the relative frequencies of the strain types among the sources and in human cases allows inference on the likely source of each infection. We introduce sourceR, an R package for quantitative source attribution, aimed at food-borne diseases. It implements a Bayesian model using strain-typed surveillance data from both human cases and source samples, capable of identifying important sources of infection. The model measures the force of infection from each source, allowing for varying survivability, pathogenicity and virulence of pathogen strains, and varying abilities of the sources to act as vehicles of infection. A Bayesian non-parametric (Dirichlet process) approach is used to cluster pathogen strain types by epidemiological behaviour, avoiding model overfitting and allowing detection of strain types associated with potentially high "virulence". sourceR is demonstrated using Campylobacter jejuni isolate data collected in New Zealand between 2005 and 2008. Chicken from a particular poultry supplier was identified as the major source of campylobacteriosis, which is qualitatively similar to results of previous studies using the same dataset. Additionally, the software identifies a cluster of 9 multilocus sequence types with abnormally high 'virulence' in humans. sourceR enables straightforward attribution of cases of zoonotic infection to putative sources of infection. As sourceR develops, we intend it to become an important and flexible resource for food-borne disease attribution studies.