A numerical taxonomic study of proteolytic and lipolytic psychrotrophs isolated from caprine milk

Lipolytic and proteolytic psychrotrophs were isolated from raw and pasteurized goats' milk, which had been stored at 5 degrees C for 7 d. The 241 strains isolated and 20 reference strains were examined by 149 biochemical, physiological, and morphological tests. The results yielded 195 characters suitable for taxonomic analysis. Computer-assisted complete linkage analysis, using the Jaccard coefficient, produced 22 phenons at 75% S. The results showed that Pseudomonas fluorescens was the predominant psychrotrophic bacterium, but that Pseudomonas fragi was dominant in some milk samples. Strains of Serratia liquefaciens and Flavobacterium balustinum were also identified.     

A numerical taxonomic study of psychrotrophic bacteria associated with lipolytic spoilage of raw milk

Raw milk samples were stored for 1-4 d and examined for bacterial growth and lipase activity. Thirty-six samples in which an increase in the heat-stable lipase activity was observed during storage were selected for further study. From these raw milk samples 205 lipolytic psychrotrophic strains were selected using butterfat agar and subsequently characterized with 86 taxonomic tests. Complete linkage cluster analysis of the taxonomic data produced two major and six minor clusters at the 83% similarity level. Pseudomonas fluorescens and Ps. fragi accounted for 63.9 and 31.2%, respectively, of the isolates.     

A numerical taxonomic study of psychrotrophic bacteria associated with lipolytic spoilage of raw milk

Raw milk samples were stored for 1–4 d and examined for bacterial growth and lipase activity. Thirty-six samples in which an increase in the heat-stable lipase activity was observed during storage were selected for further study. From these raw milk samples 205 lipolytic psychrotrophic strains were selected using butterfat agar and subsequently characterized with 86 taxonomic tests. Complete linkage cluster analysis of the taxonomic data produced two major and six minor clusters at the 83% similarity level. Pseudomonas fluorescens and Ps. fragi accounted for 63.9 and 31.2%, respectively, of the isolates.     

Application of multidishes to the characterization of proteolytic psychrotrophs from raw milk

This paper describes a procedure in which 24-well multidishes and an eight-channel pipette were used in physiological tests to characterize proteolytic, psychorotrophic and, mainly fluorescent, pseudomonads which had been isolated from refrigerated raw milk. Some of these tests have not been duplicated by any of the currently available commercial identification systems.Of the 41 biochemical tests examined, 32 could be performed in multidishes. Multidishes could be used in tests to determine carbon source utilization (27 substrates), growth on MacConkey and cetrimide agars, pigment production on King B agar, gelatin hydrolysis and levan production. Since four tests could be simultaneously inoculated, with less than 2.0 ml of medium per test, the properties of a large number of strains could be rapidly and economically determined. Differences between multidish and conventional test results in the 37°C growth test depended upon whether strains were grown on agar or in liquid medium. Hydrolyses of DNA, butterfat and egg yolk and the dissolution of crystals on tyrosine agar were more easily read when tests were performed in petri dishes than in multidishes. False positive results were obtained for melanin production on tyrosine agar. Arginine dihydrolase, Simmons' citrate and urease tests could not be performed in multidishes since the diffusion of an alkaline product, presumably ammonia, through the headspace above wells resulted in false positive reactions.     

A numerical taxonomic study of the Pseudomonas flora isolated from poultry meat

Pseudomonas strains were isolated from both fresh and cold-stored broiler skin. Phenotypically-based numerical taxonomic techniques were used to characterize the isolates and 36 reference strains. For this purpose, Biolog GN Microplates, API 20NE and a number of other biochemical tests were used. Jaccard clustering revealed the predominance of four major Pseudomonas groups: Ps. fragi, Ps. lundensis, strains belonging to Ps. fluorescens biovars and an unidentified group of strains displaying a high degree of similarity to Ps. fluorescens biovars. Within Ps. fluorescens, biovar A was best represented. The marked proteolytic character of members of Ps. fluorescens biovars A, B and C, as well as of members of the unidentified cluster, supports their possible role in the origin of organoleptic defects. In the Ps. lundensis cluster, a distinct group of Ps. lundensis-like species was found. Further genotypic studies should be carried out to clarify the taxonomic status of the Ps. lundensis-like strains and that of the unidentified group resembling Ps. fluorescens biovars A and B.     

A numerical taxonomic study of fluorescent Pseudomonas strains isolated from natural mineral waters

M. ELOMARI, L. COROLER, D. IZARD AND H. LECLERC. 1995. Forty-six strains of fluorescent pseudomonads, isolated from natural mineral waters, together with 12 strains from clinical material and 44 reference strains, were phenotypically classified by 281 characteristics. The data were processed by the Dice similarity coefficient and unweighted pair group algorithm with arithmetic averages. Eight clusters were defined at the 62% similarity level. Clusters I, II and IV were further divided into nine subclusters. Virtually all the mineral water strains fall into three groups: Ib (eight strains), Ha (14 strains) and V (16 strains). Subclusters Ib and IIa included natural mineral water strains only. Cluster V contained 13 mineral water strains and three culture collection strains of Pseudomonas fluorescens biovar III. DNA/DNA hybridization studies are needed to define the taxonomic status of these three groups within the genus Pseudomonas.     

Numerical taxonomy of proteolytic psychrotrophs from Queensland raw milks

Eighty-seven proteolytic psychrotrophic micro-organisms were isolated from 11 bulk milk supplies of two Queensland factories from different climatic regions, before and after storage at 4°C for 7 d. These isolates together with 15 reference strains formed the basis of a numerical taxonomic study involving 81 attributes. All but six isolates were pseudomonads. The strains clustered into nine groups, of which one group consisted of four yeasts. One group, containing 39 isolates, was designated as Pseudomonas fluorescens biovar 1; three groups, containing 27 isolates, as Ps. fluorescens biovar 5; and one group, containing 10 isolates, as Ps. putida biovar A. This study showed that the proteolytic psychrotrophic microflora of the 11 milks supplying the two factories was substantially different and that the proteolytic flora of 7 d refrigerated milk could not be estimated by examining the flora before storage.     

Numerical taxonomy of proteolytic psychrotrophs from Queensland raw milks

Eighty-seven proteolytic psychrotrophic micro-organisms were isolated from 11 bulk milk supplies of two Queensland factories from different climatic regions, before and after storage at 4 degrees C for 7 d. These isolates together with 15 reference strains formed the basis of a numerical taxonomic study involving 81 attributes. All but six isolates were pseudomonads. The strains clustered into nine groups, of which one group consisted of four yeasts. One group, containing 39 isolates, was designated as Pseudomonas fluorescens biovar 1; three groups, containing 27 isolates, as Ps. fluorescens biovar 5; and one group, containing 10 isolates, as Ps. putida biovar A. This study showed that the proteolytic psychrotrophic microflora of the 11 milks supplying the two factories was substantially different and that the proteolytic flora of 7 d refrigerated milk could not be estimated by examining the flora before storage.     

A numerical taxonomic study of Cucurbita

Several statistical techniques of numerical taxonomy were studied by using 21 species ofCucurbita as models. Data were taken on 93 plant characters for 24 operational taxonomic units (OTU’s). Three similarity coefficients, Q-correlation, distance and divergence, were used to compute phenetic similarities among the 24 OTU’s. The results are summarized in the form of eight phenograms using both unweighted and weighted pair-group methods of clustering. The results obtained from the various statistical techniques were compared with cross-compatibility ratings. The distance and divergence coefficients and their respective phenogram values were more highly correlated with the cross-compatibility ratings than were the Qcorrelation coefficients and their phenogram values. However, from a subjective viewpoint, the phenograms derived from Q-correlation coefficients were in closer general agreement with a combination of biological factors including cross compatibility, geographical distribution, and ecological adaption than were phenograms derived from distance or divergence coefficients.     

A numerical taxonomic study of Pseudomonas strains from spoiled meat

A numerical taxonomic study using 160 unit characters has been performed on 110 isolates of pseudomonads from aerobically spoiled beef and pork and on 13 named strains from the genus Pseudomonas. Four clusters of meat strains were detected at 87% S. Non-fluorescent strains were contained in two closely related clusters (1 and 2) which were identified with P. fragi. Clusters 3 and 4 contained fluorescent strains which were distinct from P. fluorescens, P. putida and the other named strains examined. An identification scheme based on 11 carbon source tests is presented for the recognition of cluster members.     

A numerical taxonomic study of Flavobacterium-Cytophaga strains from dairy sources

Phenotypic data on 203 Gram-negative non-fermentative bacteria of the Flavobacterium-Cytophaga group isolated from milk and butter were analyzed by numerical taxonomic techniques. Twenty reference strains including species of Flavobacterium, Cytophaga and strains of Pseudomonas paucimobilis were included in the study. Using the matching coefficient of Sokal & Michener with antibiotic susceptibility data included, 189 isolates were recovered in nine clusters. Six of these clusters were linked at or above the 85% S level while three were linked at or above the 79% S level. The largest cluster, representing 46.3% of the isolates, could be equated with Flavobacterium sp. Group IIb. Other clusters could be equated with Flavobacterium sp. L 16/1 (22.7% of isolates), F. balustinum (10.8% of isolates), F. breve (4.4%), F. multivorum (3.5%) and Cytophaga johnsonae (1.5%). The cluster resembling Flavobacterium sp. L 16/1 and a smaller unclassified cluster, were exceptional in being susceptible to the antibiotics cephalothin and penicillin G.     

A numerical taxonomic study of Flavobacterium-Cytophaga strains from dairy sources

Phenotypic data on 203 Gram-negative non-fermentative bacteria of the Flavobacterium-Cytophaga group isolated from milk and butter were analyzed by numerical taxonomic techniques. Twenty reference strains including species of Flavobacterium, Cytophaga and strains of Pseudomonas paucimobilis were included in the study. Using the matching coefficient of Sokal & Michener with antibiotic susceptibility data included, 139 isolates were recovered in nine clusters. Six of these clusters were linked at or above the 85% S level while three were linked at or above the 79% S level. The largest cluster, representing 46.3% of the isolates, could be equated with Flavobacterium sp. Group IIb. Other clusters could be equated with Flavobacterium sp. L 16/1 (22.7% of isolates), F. balustinum (10.8% of isolates), F. breve (4.4%), F. multivorum (3.5%) and Cytophaga johnsonae (1.5%). The cluster resembling Flavobacterium sp. L 16/1 and a smaller unclassified cluster, were exceptional in being susceptible to the antibiotics cephalothin and penicillin G.     

A numerical taxonomic study of Propionibacterium strains from dairy sources

T.J. BRITZ AND K-H.J. RIEDEL. 1991. Phenotypic results from 81 tests conducted on 73 propionibacteria, including five type strains, 22 reference strains, unidentified propionibacteria and strains isolated from dairy sources, were analysed by numerical taxonomy. Characters giving uniform results were excluded. With the simple matching coefficient and the single linkage cluster analysis, 61 cultures were recovered in five major clusters. Final linkage of all the Propionibacterium cultures examined was at the 77% S-level. The results clearly showed that it was possible to distinguish between the 'classical' and 'cutaneous' Propionibacterium spp., corresponding with the type habitat of each group. The major 'classical' clusters were equated with the P. freudenreichii, P. thoenii, P. jensenii and P. acidipropionici species, while the only major 'cutaneous' cluster was equated with the P. acnes species. The major clusters were identified by relating them to specific type strains and by comparing phenotypic characteristics. The differentiating characteristics of each cluster were determined. The largest cluster, representing 37% of the strains, was equated with P. jensenii but contained cultures that produced an atypical brown/red pigment. These strains, although positively identified as P. jensenii , could also be identified as the 'old' P. rubrum ' species. Thus if pigmentation is used as differential characteristic two distinct groups of propionibacteria could be identified within the P. jensenii species.     

A numerical taxonomic study of Leuconostoc oenos strains from wine

Phenotypic data of 108 tests conducted on 70 malolactic bacteria, including 54 presumptive Leuconostoc oenos strains, five presumptive pediococci isolated from wine and 11 reference strains, were analysed by numerical taxonomic techniques. Using the simple matching coefficient, 58 strains were grouped in six clusters at the 87% S level. Cell wall analysis of the interpeptide bridges and morphology was also used to differentiate between the strains. No L. oenos reference strains were found to group in any cluster. The hypothetical median organism (HMO) of cluster A was related at only the 63% S level to the L. oenos type strain and it is proposed that these strains be regarded as 'L. gracile'. The majority of the L. oenos strains (35) grouped together at above the 91% S level in cluster B, with the HMO of this cluster related at the 75% S level to the L. oenos type strain. Results indicate that the majority of L. oenos strains included in this study are closely related, but more research is needed to justify the separation of these strains into more than one species.     

A numerical taxonomic study of some pigmented bacteria isolated from Organic Lake,an antarctic hypersaline lake

A study was made of a group of moderately halophilic, heterotrophic, pigmented strains isolated from Organic Lake, Antarctica. These strains were Gram-negative, non-motile, had an aerobic metabolism and a mol% G+C content of their DNA in the range 35–41, indicating that they may be members of the Flavobacterium-Cytophaga group. A numerical taxonomic study involving 134 characteristics compared the antarctic strains with reference strains from Flavobacterium, Cytophaga and Flectobacillus. The antarctic strains formed two clusters that did not contain any reference strains suggesting that they may represent two new species of the genus Flavobacterium.     

A numerical taxonomic study of species of Vibrio isolated from the aquatic environment and birds in Kent, England

A numerical taxonomic study has been carried out to confirm the identity of strains of the family Vibrionaceae isolated during an ecological study. A total of 237 strains were studied including 148 from the aquatic environment, 6 from estuarine birds, 1 from sheep faeces, and 61 control cultures. Duplicates of 21 of the strains were randomly selected and included to estimate test and operator error. Taxonomic resemblance was estimated on the basis of 148 characters using Euclidean distance. The taxonomic position of some strains was reevaluated using the pattern different coefficient. Strains were clustered by three methods, all of which gave similar results. The estimated average probability of test error was 1.5%. Strains previously identified as Vibrio anguillarum fell into four distinct phenons corresponding to V. anguillarum biovar I, 'V. anguillarum biovar II', V. diazotrophicus, and strains pathogenic to oyster larvae. The latter group characteristically degraded xanthine and probably represents a new species. The phenon corresponding to V. cholerae included the type strain, strains of human origin, and strains isolated in the United Kingdom from birds and the aquatic environment. Some strains of V. cholerae were luminous. Other phenons were identified as V. metschnikovii, V. fluvialis, and Aeromonas spp.     

A numerical taxonomic study on heterophyid trematodes

A numerical taxonomy was studied on a group of heterophyid trematodes and analysis was made on the following species: Metagonimus yokogawai (3 OTU, Operational Taxonomic Unit), Metagonimus Miyata Type (3 OTU), Metagonimus takahashii (2 OTU), Heterophyes dispar (2 OTU), Heterophyes heterophyes (1 OTU), Heterophyes nocens (2 OTU), Heterophyopsis continua (1 OTU), Pygidiopsis summa (3 OTU), Stellantchasmus falcatus (2 OTU) and Stictodora lari (2 OTU). Twenty-six morphological characters were measured and their values were expressed as relative ratios. Similarity and correlation matrix among each individuals were calculated. Clustering analysis by Ward's method and factor analysis were performed using the SAS (Statistical Analysis System) package. As a result, the groups belonging to the genus of Metagonimus were divided into three phenons (Metagonimus yokogawai, Metagonimus Miyata Type, M. takahashii), and Metagonimus Miyata Type was classified as the level of subspecies of M. takahashii. The groups belonging to the genus Heterophyes were clearly divided into three phenons (Heterophyes dispar, H. heterophyes, H. nocens), and H. nocens was classified as not a subspecies level of H. heterophyes but a distinct species. Other species were classified as distinct phenons. From these results, the application of numerical taxonomy on trematode classification is considered to be a great aid to determine the limit of taxa.     

A numerical taxonomic study of species of Vibrio isolated from the aquatic environment and birds in Kent, England

A numerical taxonomic study has been carried out to confirm the identity of strains of the family Vibrionaceae isolated during an ecological study. A total of 237 strains were studied including 148 from the aquatic environment, 6 from estuarine birds, 1 from sheep faeces, and 61 control cultures. Duplicates of 21 of the strains were randomly selected and included to estimate test and operator error. Taxonomic resemblance was estimated on the basis of 148 characters using Euclidean distance. The taxonomic position of some strains was reevaluated using the pattern difference coefficient. Strains were clustered by three methods, all of which gave similar results. The estimated average probability of test error was 1.5%. Strains previously identified as Vibrio anguillarum fell into four distinct phenons corresponding to V. anguillarum biovar I, ' V. anguillarum biovar II', V. diazotrophicus , and strains pathogenic to oyster larvae. The latter group characteristically degraded xanthine and probably represents a new species. The phenon corresponding to V. cholerae included the type strain, strains of human origin, and strains isolated in the United Kingdom from birds and the aquatic environment. Some strains of V. cholerae were luminous. Other phenons were identified as V. metschnikovii, V. fluvialis , and Aeromonas spp.     

A numerical taxonomic study of non-motile non-fermentative Gram-negative bacteria from foods

A numerical taxonomic study using 155 unit characters was performed on 63 strains of Gram-negative non-motile non-fermentative bacteria isolated from proteinaceous foods. Similar bacteria from other sources and several Pseudomonas strains from meat were included for reference purposes. Three clusters were observed at 76% S which contained all the food strains. Cluster 1 was composed entirely of Acinetobacter strains including 17 isolated from foods that were provisionally identified with Acinetobacter johnsonii. Cluster 2 contained 22 strains identified as Psychrobacter immobilis , of which 20 were from food. Cluster 3 contained all the Pseudomonas reference strains and 26 non-motile strains isolated from meat. These were shown to be non-motile variants of Pseudomonas fragi. A simple identification scheme, based on five tests, is presented for the distinction of the three types of bacteria.