BACKGROUND: Terminal restriction fragment length polymorphism (T-RFLP) analysis is a common DNA-fingerprinting technique used for comparisons of complex microbial communities. Although the technique is well established there is no consensus on how to treat T-RFLP data to achieve the highest possible accuracy and reproducibility. This study focused on two critical steps in the T-RFLP data treatment: the alignment of the terminal restriction fragments (T-RFs), which enables comparisons of samples, and the normalization of T-RF profiles, which adjusts for differences in signal strength, total fluorescence, between samples. RESULTS: Variations in the estimation of T-RF sizes were observed and these variations were found to affect the alignment of the T-RFs. A novel method was developed which improved the alignment by adjusting for systematic shifts in the T-RF size estimations between the T-RF profiles. Differences in total fluorescence were shown to be caused by differences in sample concentration and by the gel loading. Five normalization methods were evaluated and the total fluorescence normalization procedure based on peak height data was found to increase the similarity between replicate profiles the most. A high peak detection threshold, alignment correction, normalization and the use of consensus profiles instead of single profiles increased the similarity of replicate T-RF profiles, i.e. lead to an increased reproducibility. The impact of different treatment methods on the outcome of subsequent analyses of T-RFLP data was evaluated using a dataset from a longitudinal study of the bacterial community in an activated sludge wastewater treatment plant. Whether the alignment was corrected or not and if and how the T-RF profiles were normalized had a substantial impact on ordination analyses, assessments of bacterial dynamics and analyses of correlations with environmental parameters. CONCLUSIONS: A novel method for the evaluation and correction of the alignment of T-RF profiles was shown to reduce the uncertainty and ambiguity in alignments of T-RF profiles. Large differences in the outcome of assessments of bacterial community structure and dynamics were observed between different alignment and normalization methods. The results of this study can therefore be of value when considering what methods to use in the analysis of T-RFLP data.

BACKGROUND: Terminal restriction fragment length polymorphism (T-RFLP) analysis is a DNA-fingerprinting method that can be used for comparisons of the microbial community composition in a large number of samples. There is no consensus on how T-RFLP data should be treated and analyzed before comparisons between samples are made, and several different approaches have been proposed in the literature. The analysis of T-RFLP data can be cumbersome and time-consuming, and for large datasets manual data analysis is not feasible. The currently available tools for automated T-RFLP analysis, although valuable, offer little flexibility, and few, if any, options regarding what methods to use. To enable comparisons and combinations of different data treatment methods an analysis template and an extensive collection of macros for T-RFLP data analysis using Microsoft Excel were developed. RESULTS: The Tools for T-RFLP data analysis template provides procedures for the analysis of large T-RFLP datasets including application of a noise baseline threshold and setting of the analysis range, normalization and alignment of replicate profiles, generation of consensus profiles, normalization and alignment of consensus profiles and final analysis of the samples including calculation of association coefficients and diversity index. The procedures are designed so that in all analysis steps, from the initial preparation of the data to the final comparison of the samples, there are various different options available. The parameters regarding analysis range, noise baseline, T-RF alignment and generation of consensus profiles are all given by the user and several different methods are available for normalization of the T-RF profiles. In each step, the user can also choose to base the calculations on either peak height data or peak area data. CONCLUSIONS: The Tools for T-RFLP data analysis template enables an objective and flexible analysis of large T-RFLP datasets in a widely used spreadsheet application.

MOTIVATION: Analyses of genomic signatures are gaining attention as they allow studies of species-specific relationships without involving alignments of homologous sequences. A naïve Bayesian classifier was built to discriminate between different bacterial compositions of short oligomers, also known as DNA words. The classifier has proven successful in identifying foreign genes in Neisseria meningitis. In this study we extend the classifier approach using either a fixed higher order Markov model (Mk) or a variable length Markov model (VLMk). RESULTS: We propose a simple algorithm to lock a variable length Markov model to a certain number of parameters and show that the use of Markov models greatly increases the flexibility and accuracy in prediction to that of a naïve model. We also test the integrity of classifiers in terms of false-negatives and give estimates of the minimal sizes of training data. We end the report by proposing a method to reject a false hypothesis of horizontal gene transfer. AVAILABILITY: Software and Supplementary information available at www.cs.chalmers.se/~dalevi/genetic_sign_classifiers/.