Computer-assisted statistical analyses of enzyme and ribosomal DNA electrophoretic polymorphism in Yersinia

Application of a quantitative algorithm to restriction endonuclease analysis of Aujeszky's disease (pseudorabies) virus from a geographically localized outbreak

A retrospective epidemiologic study was conducted to evaluate the application of an objective quantitative algorithm for estimating genetic similarity from restriction endonuclease analysis data. The analysis was performed to assist the determination of chronologic trends in an Aujeszky's disease viral epidemic in a geographic region. DNA from each viral isolate obtained during the epidemic was digested with 4 restriction endonucleases and molar ratio labeled to generate separate fragment patterns that were simultaneously compared using the algorithm. The resultant estimates of genetic similarity were then used in conjunction with time of virus isolation and specific geographic location of the outbreaks to identify the probable sources of infection and the patterns of spread among swine herds. This type of quantitative analysis enabled a more precise and objective approach than previously has been applied to the interpretation of restriction endonuclease data, thereby demonstrating the benefit of this methodology for the investigation of infectious disease outbreaks.

Quantitative assessment of genomic similarity from restriction fragment patterns

A quantitative algorithm for comparing restriction fragment patterns (RFPs) was developed and used to estimate the genomic similarity of 18 isolates of pseudorabies virus of known origin. Variants of this algorithm using either untransformed or square-root-transformed differences between fragment sizes were investigated with regard to their ability to classify RFPs. Multidimensional scaling was used to represent spatially the genomic relatedness among samples, with 3 dimensions producing the most meaningful results. The square-root transformation provided more interpretable dimensions. The first dimension differentiated samples geographically, separating North American from European isolates. The second and third dimensions differentiated isolates with specific gene deletions (gE and gG, respectively) from those not having these deletions. Clusters of isolates were identified that were related either by collection from the same geographic area during a specific time period, or by laboratory intervention to create vaccines. These methods offer increased precision in the determination of genetic relatedness based on RFPs, and thus offer increased diagnostic accuracy for the determination of sources of infection.

From image processing to classification: II. Classification of electrophoretic patterns using self-organizing feature maps and feed-forward neural networks

In a recent study, isoelectric focusing patterns were classified with a neural network using the back-propagation algorithm [1]. In order to further study the classification process and to generalize the presentation of electrophoretic patterns, Kohonen's self-organizing feature maps [2] were applied in this study. Although these feature maps are very efficient in many pattern recognition tasks, our data proved to be too complex for classification with an unsupervised system. Therefore, a second supervised network on top of the feature map was necessary. As in [3], a feed-forward network trained by the back-propagation algorithm was used. The final system allows us to correctly classify 90% of all wheat varieties. Moreover, the system proved to be reliable, reasonable in training time and shows the same accuracy in different experimental setups.

Genetic diversity of Pseudomonas aeruginosa strains isolated from patients with cystic fibrosis revealed by restriction fragment length polymorphism of the rRNA gene region

The restriction fragment length polymorphism patterns of rDNAs from Pseudomonas aeruginosa strains isolated from the respiratory tracts of patients suffering from cystic fibrosis were obtained to evaluate the genetic polymorphism of this population of strains. Eighty-seven P. aeruginosa strains isolated from 87 patients from diverse areas of France and the ATCC 10145 strain were examined. Four restriction enzymes were used: BamHI, ClaI, EcoRI, and PstI. Forty-nine strains (56%) were in the three most frequent ribotypes (ribotypes R1 to R3). In addition, hierarchical clustering analysis of the data showed that 71 of the 88 strains (81%) clustered at a distance of less than one-third of the greatest distance observed in the total population. This indicates that clinical isolates implicated in the pathology of cystic fibrosis present a low degree of heterogeneity of rRNA operons, in contrast to the heterogeneity of strains of P. aeruginosa isolated from patients with various other pathologies. This relative homogeneity of rRNA genes was observed independently of the clinical status of the patient and the colony morphology.

Classification of wheat varieties by isoelectric focusing patterns of gliadins and neural network

Classification of wheat varieties, using isoelectric focusing patterns of the gliadins, image processing and neural networks, is described. The method was compared to a statistical classification method, discriminant analysis. The isoelectric point and the area of each band were calculated by image processing. Different methods of presenting the electrophoretic patterns to the neural network were studied. The most effective method was transformation of the electrophoretic pattern to a small (11 x 47 pixels) representation of the original digitized image, which was presented to the neural network as a vector. The neural network was trained with a number of patterns and tested with new patterns from different electrophoretic runs of the same wheat varieties. In this study we used ten different wheat varieties and the neural network was able to classify 95.5% of the patterns correctly. The statistical classification method classified the same data set 91.8% correctly. We conclude that both the neural network and discriminant analysis were able to classify the patterns correctly with a high degree of certainty. The patterns that were misclassified were indistinguishable by visual inspection.

The absence of correlation between allozyme and rrn RFLP analysis indicates a high gene flow rate within human clinical Pseudomonas aeruginosa isolates

Two hundred and fifty seven human clinical Pseudomonas aeruginosa strains isolated between 1984 and 1990 in several regions of France, as well as two reference strains, were studied by computer-assisted statistical analysis of the data from their esterase electrophoretic patterns and rrn restriction fragment length polymorphisms. No correlation was found between the two sets of data except for some strains of serotype O12 which, thus, may constitute a distinct group within the species. This absence of correlation indicates a high gene flow rate within human isolates of the P. aeruginosa species. A possible explanation is that, because of an as yet unidentified selective advantage, the esterase loci are a major target for recombinational events. Alternatively, horizontal genetic transfers between strains may have occurred at so high a rate that the clonal structure usually observed in bacterial populations has been disrupted. This study highlights clearly the need for caution in inferring bacterial population structure from any single class of genetic markers.