Artificial neural network analysis of pyrolysis mass spectrometric data in the identification of Streptomyces strains

Abstract Sixteen representatives of three morphologically distinct groups of streptomycetes were recovered from soil using selective isolation procedures. Duplicated batches of the test strains were examined by Curie-point pyrolysis mass spectrometry and the first data set used for conventional multivariate statistical analyses and as a training set for an artificial neural network. The second set of data was used for 'operational fingerprinting' and for testing the artificial neural network. All of the test strains were correctly identified using the artificial neural network whereas only fifteen of the sixteen strains were assigned to the correct group using the conventional operational fingerprinting procedure. Artificial neural network analysis of pyrolysis mass spectrometric data provides a rapid, cost-effective and reproducible way of identifying and typing large numbers of microorganisms.     

西双版纳、白茫雪山、波罗的海南岸地衣纯培养放线菌多样性

【背景】在地衣共生系统中除了共生真菌、藻类以外,还蕴藏着丰富的放线菌资源。【目的】采集来自云南西双版纳、白茫雪山、德国波罗的海南岸3个地区的地衣,对获得的地衣纯培养放线菌进行多样性分析。【方法】采用3种放线菌选择分离培养基,通过平板稀释涂布法分离放线菌。通过比较16S rRNA基因序列相似性以及构建系统发育树,确定纯培养放线菌的分类地位。【结果】共分离菌株1 123株,鉴定417株。其中从西双版纳17份地衣样品中分离纯化到107株放线菌,分布在7个目14个科33个属,潜在新种18株,其中链霉菌为优势菌属;从白茫雪山7份地衣样品中分离纯化到103株放线菌,分布在4个目5个科9个属,潜在新种16株。其中链霉菌为优势菌属,占比39%;从波罗的海南岸5份地衣样品中分离纯化到65株放线菌,分布在4个目8个科18个属,潜在新种5株,潜在新种菌和链霉菌为优势菌属。【结论】在本研究条件下,西双版纳可培养地衣放线菌多样性较白茫雪山和波罗的海南岸丰富。白茫雪山地衣链霉菌居多,潜在新种占比15.5%。3个地区地衣放线菌的区系组成各不相同,这与3个地区地衣所处地理环境、完全不同的气候密切相关。     

Identification of some industrially importantActinoplanes species

Summary Three actinomycetes designated A294, A385 and A448 all produced novel inhibitors of an enzyme involved in the regulation of mammalian cholesterol metabolism. The organisms were identified as members of the genusActinoplanes using morphological and chemotaxonomic characters. Numerical phenetic data obtained for clusters defined at the 77.5% S_sm similarity level in the taxonomic study ofActinoplanes and related taxe by Goodfellow et al. [6], was used to construct a probabilistic identification matrix for the genusActinoplanes. This was in an attempt to characterise the producing strains to species level. Identification scores for the three organisms, together with two type strains, were determined using the MATIDEN program of Sneath [36]. Strains A294 and A385 did not identify correctly with any cluster. Strain A448 was identified asActinoplanes derwentensis. The two type strains identified within acceptable statistical limits to their correct clusters. Due to the absence of other taxonomic criteria with which to speciate members of the genusActinoplanes we recommend that the two unidentified producing strains should be considered novel until such time as surther techniques become available.     

The deconvolution of pyrolysis mass spectra using genetic programming: application to the identification of some Eubacterium species

Pyrolysis mass spectrometry was used to produce complex biochemical fingerprints of Eubacterium exiguum, E. infirmum, E. tardum and E. timidum. To examine the relationship between these organisms the spectra were clustered by canonical variates analysis, and four clusters, one for each species, were observed. In an earlier study we trained artificial neural networks to identify these clinical isolates successfully; however, the information used by the neural network was not accessible from this so-called `black box' technique. To allow the deconvolution of such complex spectra (in terms of which masses were important for discrimination) it was necessary to develop a system that itself produces `rules' that are readily comprehensible. We here exploit the evolutionary computational technique of genetic programming; this rapidly and automatically produced simple mathematical functions that were also able to classify organisms to each of the four bacterial groups correctly and unambiguously. Since the rules used only a very limited set of masses, from a search space some 50 orders of magnitude greater than the dimensionality actually necessary, visual discrimination of the organisms on the basis of these spectral masses alone was also then possible.     

Identification and Discrimination of Oral Asaccharolytic Eubacterium spp. by Pyrolysis Mass Spectrometry and Artificial Neural Networks

Curie-point pyrolysis mass spectra were obtained from 29 oral asaccharolytic Eubacterium strains and 6 abscess isolates previously identified as Peptostreptococcus heliotrinreducens. Pyrolysis mass spectrometry (PyMS) with cluster analysis was able to clarify the taxonomic position of this group of organisms. Artificial neural networks (ANNs) were then trained by supervised learning (with the back-propagation algorithm) to recognize the strains from their pyrolysis mass spectra; all Eubacterium strains were correctly identified, and the abscess isolates were identified as un-named Eubacterium taxon C₂ and were distinct from the type strain of P. heliotrinreducens. These results demonstrate that the combination of PyMS and ANNs provides a rapid and accurate identification technique.     

Marine actinomycetes as a source of novel secondary metabolites

A set of 600 actinomycetes strains which were isolated from marine sediments from various sites in the Pacific and Atlantic Oceans were screened for the production of bioactive secondary metabolites. Marine streptomycete strains were found to be producers of well known chemically diverse antibiotics isolated from terrestrial streptomycetes, as in the case of marine Micromonospora strains. New marine members of the rare genus Verrucosispora seem to be a promising source for novel bioactive secondary metabolites as shown in the case of the abyssomicin producing strain AB-18-032.     

Artificial neural network based identification of Campylobacter species by Fourier transform infrared spectroscopy

Two prototypes of artificial neural network (ANN), multilayer perceptron (MLP), and probabilistic neural network (PNN), were used to analyze infrared (IR) spectral data obtained from intact cells belonging to the species Campylobacter coli and Campylobacter jejuni. In order to establish a consistent identification and typing procedure, mid infrared spectra of these species were obtained by means of a Fourier transform infrared (FT-IR) spectroscope. FT-IR patterns belonging to 26 isolates subclassified into 4 genotypes were pre-processed (normalized, smoothed and derivatized) and grouped into training, verification and test sets. The two architectures tested (PNN, MLP) were developed and trained to identify or leave unassigned a number of IR patterns. Two window ranges (w(4), 1200 to 900 cm(-1); and w(5), 900 to 700 cm(-1)) in the mid IR spectrum were presented as input to the ANN models functioning as pattern recognition systems. No matter the ANN used all the training sets were correctly identified at subspecies level. For the test set, the four-layer MLP network was found to be specially suitable to recognize FT-IR data since it correctly identified 99.16% of unknowns using the w(4) range, and was fully successful in detecting atypical patterns from closely related Campylobacter strains and other bacterial species. The PNN network obtained lower percentages in assignation and rejection. Overall, ANNs constitute an excellent mathematical tool in microbial identification, since they are able to recognize with a high degree of confidence typical as well as atypical FT-IR fingerprints from Campylobacter spp.     

Biosystematics and diversity amongst novel carboxydotrophic actinomycetes

Fifty-four carboxydotrophic actinomycetes isolated from soils and composts were compared through 119 unit characters with representative mesophilic and thermophilic streptomycetes. The data were examined using the Jaccard, pattern and simple matching coefficients and clustering achieved using the unweighted pair group method with arithmetic averages algorithm. Acceptable cophenetic correlation and test error values allowed confidence to be placed in the resultant numerical taxonomies. The carboxydotrophic actinomycetes, which were distinct from cluster-groups corresponding to the mesophilic and thermophilic streptomycetes, formed two major cluster-groups the members of which were examined for the presence of diagnostic chemical markers. All but two of the carboxydotrophic actinomycetes had a profile of chemical properties consistent with their assignment to the genusStreptomyces. Quantitative fatty acid data were examined using the SIMCA package and the two statistically significant groups obtained corresponded with the cluster-groups circumscribed in the numerical phenetic analysis. Members of the two groups were also distinguished on the basis of their phospholipid composition. The two strains that containedmeso- as opposed to LL-diaminopimelic acid in their peptidoglycan also showed a distinct chemotaxonomic profile. It was concluded that the carboxydotrophic actinomycetes form a novel and taxonomically diverse group.     

Pyrolysis patterns of 5 close Corynebacterium species analyzed by artificial neural networks

In the present study, an artificial neural network was trained with the Stuttgart Neural Networks Simulator, in order to identify Corynebacterium species by analyzing their pyrolysis patterns. An earlier study described the combination of pyrolysis, gas chromatography and atomic emission detection we used on whole cell bacteria. Carbon, sulfur and nitrogen were detected in the pyrolysis compounds. Pyrolysis patterns were obtained from 52 Corynebacterium strains belonging to 5 close species. These data were previously analyzed by Euclidean distances calculation followed by Unweighted Pair Group Method of Averages, a clustering method. With this early method, strains from 3 of the 5 species (C. xerosis, C. freneyi and C. amycolatum) were correctly characterized even if the 29 strains of C. amycolatum were grouped into 2 subgroups. Strains from the 2 remaining species (C. minutissimum and C. striatum) cannot be separated. To build an artificial neural network, able to discriminate the 5 previous species, the pyrolysis data of 42 selected strains were used as learning set and the 10 remaining strains as testing set. The chosen learning algorithm was Back-Propagation with Momentum. Parameters used to train a correct network are described here, and the results analyzed. The obtained artificial neural network has the following cone-shaped structure: 144 nodes in input, 25 and 9 nodes in 2 successive hidden layers, and then 5 outputs. It could classify all the strains in their species group. This network completes a chemotaxonomic method for Corynebacterium identification.     

A Model for Improving the Learning Curves of Artificial Neural Networks

In this article, the performance of a hybrid artificial neural network (i.e. scale-free and small-world) was analyzed and its learning curve compared to three other topologies: random, scale-free and small-world, as well as to the chemotaxis neural network of the nematode Caenorhabditis Elegans. One hundred equivalent networks (same number of vertices and average degree) for each topology were generated and each was trained for one thousand epochs. After comparing the mean learning curves of each network topology with the C. elegans neural network, we found that the networks that exhibited preferential attachment exhibited the best learning curves.     

Isolation of four aquatic streptomycetes strains capable of growth on organochlorine pesticides

Ninety-three wild-type isolates identified as actinomycetes were tested against 11 organochlorine pesticides (OPs): aldrin, chlordane, DDD, DDE, DDT, dieldrin, heptachlor, and heptachlor epoxides, lindane, and methoxychlor. Qualitative screening agar assays displayed 62-78% tolerance of strains to OPs. Four strains designed M4, M7, M9 and M15 were selected based on multi-OP-tolerance, and identified as members of the streptomycetes group. Different growth profiles were observed in cultures of the four selected streptomycetes cultured in synthetic medium containing 5-50 microg x l(-1) aldrin or chlordane or lindane. Increase of aldrin removal by the selected microorganisms was concomitant with the 4.8-36.0 microg x l(-1) pesticide concentration range. After 72 h of streptomycete M7 growth in synthetic medium containing 48.0 microg x l(-1) aldrin, the remaining OP concentration in the supernatant was approximately 10% of the initial concentration. Also, in stationary growth phase less than 2.5 microg x l(-1) aldrin residual concentration was detected in the medium.     

Using a neural network and spatial clustering to predict the location of active sites in enzymes

Structural genomics projects aim to provide a sharp increase in the number of structures of functionally unannotated, and largely unstudied, proteins. Algorithms and tools capable of deriving information about the nature, and location, of functional sites within a structure are increasingly useful therefore. Here, a neural network is trained to identify the catalytic residues found in enzymes, based on an analysis of the structure and sequence. The neural network output, and spatial clustering of the highly scoring residues are then used to predict the location of the active site.A comparison of the performance of differently trained neural networks is presented that shows how information from sequence and structure come together to improve the prediction accuracy of the network. Spatial clustering of the network results provides a reliable way of finding likely active sites. In over 69% of the test cases the active site is correctly predicted, and a further 25% are partially correctly predicted. The failures are generally due to the poor quality of the automatically generated sequence alignments.We also present predictions identifying the active site, and potential functional residues in five recently solved enzyme structures, not used in developing the method. The method correctly identifies the putative active site in each case. In most cases the likely functional residues are identified correctly, as well as some potentially novel functional groups.     

Selection and Characterization of Microorganisms Utilizing Thaxtomin A, a Phytotoxin Produced by Streptomyces scabies

Thaxtomin A is the main phytotoxin produced by Streptomyces scabies, a causal agent of potato scab. Thaxtomin A is a yellow compound composed of 4-nitroindol-3-yl-containing 2,5-dioxopiperazine. A collection of nonpathogenic streptomycetes isolated from potato tubers and microorganisms recovered from a thaxtomin A solution were examined for the ability to grow in the presence of thaxtomin A as a sole carbon or nitrogen source. Three bacterial isolates and two fungal isolates grew in thaxtomin A-containing media. Growth of these organisms resulted in decreases in the optical densities at 400 nm of culture supernatants and in 10% reductions in the thaxtomin A concentration. The fungal isolates were identified as a Penicillium sp. isolate and a Trichoderma sp. isolate. One bacterial isolate was associated with the species Ralstonia pickettii, and the two other bacterial isolates were identified as Streptomyces sp. strains. The sequences of the 16S rRNA genes were determined in order to compare thaxtomin A-utilizing actinomycetes to the pathogenic organism S. scabies and other Streptomyces species. The nucleotide sequences of the γ variable regions of the 16S ribosomal DNA of both thaxtomin A-utilizing actinomycetes were identical to the sequence of Streptomyces mirabilis ATCC 27447. When inoculated onto potato tubers, the three thaxtomin A-utilizing bacteria protected growing plants against common scab, but the fungal isolates did not have any protective effect.     

Biosystematic studies on novel streptomycetes from soil

Members of three putatively novel Streptomyces species, designated Streptomyces groups A, B and C, were repeatedly isolated from environmental samples taken from four hay meadow plots at Cockle Park Experimental Farm, Northumberland (UK). Representative isolates were found to have properties consistent with their classification in the genus Streptomyces and were recovered in three taxa using different phenotypic criteria, namely morphological and pigmentation properties, rapid enzyme tests, and whole-organism fatty acid, protein electrophoretic and pyrolysis mass-spectrometric data. The isolates were rapidly characterised as three taxonomic groups using pyrolysis mass spectrometry. The three taxa were also distinguished from one another and from validly described species of Streptomyces using rapid enzyme tests based on the fluorophores 7-amino-methylcoumarin and 4-methylumbelliferone, and computer-assisted identification procedures. The results indicate that selective isolation and rapid characterisation of streptomycetes using pyrolysis mass spectrometry provide a practical way of determining the phenotypic species diversity of streptomycetes in natural habitats. The experimental data also indicate that representative sampling of cultivable streptomycetes from soil can best be achieved using a multi-step extraction procedure coupled with the use of selective isolation procedures.     

Marine actinomycetes: An ongoing source of novel bioactive metabolites

Actinomycetes are virtually unlimited sources of novel compounds with many therapeutic applications and hold a prominent position due to their diversity and proven ability to produce novel bioactive compounds. There are more than 22,000 known microbial secondary metabolites, 70% of which are produced by actinomycetes, 20% from fungi, 7% from Bacillus spp. and 1–2% by other bacteria. Among the actinomycetes, streptomycetes group are considered economically important because out of the approximately more than 10,000 known antibiotics, 50–55% are produced by this genus. The ecological role of actinomycetes in the marine ecosystem is largely neglected and various assumptions meant there was little incentive to isolate marine strains for search and discovery of new drugs. The search for and discovery of rare and new actinomycetes is of significant interest to drug discovery due to a growing need for the development of new and potent therapeutic agents. Modern molecular technologies are adding strength to the target-directed search for detection and isolation of bioactive actinomycetes, and continued development of improved cultivation methods and molecular technologies for accessing the marine environment promises to provide access to this significant new source of chemical diversity with novel/rare actinomycetes including new species of previously reported actinomycetes.     

Rapid identification using pyrolysis mass spectrometry and artificial neural networks of Propionibacterium acnes isolated from dogs

Curie-point pyrolysis mass spectra were obtained from reference Propionibacterium strains and canine isolates. Artificial neural networks (ANNs) were trained by supervised learning (with the back-propagation algorithm) to recognize these strains from their pyrolysis mass spectra; all the strains isolated from dogs were identified as human wild type P. acnes. This is an important nosological discovery, and demonstrates that the combination of pyrolysis mass spectrometry and ANNs provides an objective, rapid and accurate identification technique. Bacteria isolated from different biopsy specimens from the same dog were found to be separate strains of P. acnes , demonstrating a within-animal variation in microflora. The classification of the canine isolates by Kohonen artificial neural networks (KANNs) was compared with the classical multivariate techniques of canonical variates analysis and hierarchical cluster analysis, and found to give similar results. This is the first demonstration, within microbiology, of KANNs as an unsupervised clustering technique which has the potential to group pyrolysis mass spectra both automatically and relatively objectively.     

Distinction of males of the Lutzomyia intermedia (Lutz & Neiva, 1912) species complex by ratios between dimensions and by an artificial neural network (Diptera: Psychodidae, Phlebotominae)

The females of the two species of the Lutzomyia intermedia complex can be easily distinguished, but the males of each species are quite similar. The ratios between the extra-genital and the genital structures of L. neivai are larger than those of L. intermedia s. s., according to ANOVA. An artificial neural network was trained with a set of 300 examples, randomly taken from a sample of 358 individuals. The input vectors consisted of several ratios between some structures of each insect. The model was tested on the remaining 58 insects, 56 of which (96.6%) were correctly identified. This ratio of success can be considered remarkable if one takes into account the difficulty of attaining comparable results using traditional statistical techniques.     

Differentiation of Listeria monocytogenes serovars by using artificial neural network analysis of Fourier-transformed infrared spectra

A classification system based on Fourier transform infrared (FTIR) spectroscopy combined with artificial neural network analysis was designed to differentiate 12 serovars of Listeria monocytogenes using a reference database of 106 well-defined strains. External validation was performed using a test set of another 166 L. monocytogenes strains. The O antigens (serogroup) of 164 strains (98.8%) could be identified correctly, and H antigens were correctly determined in 152 (91.6%) of the test strains. Importantly, 40 out of 41 potentially epidemic serovar 4b strains were unambiguously identified. FTIR analysis is superior to PCR-based systems for serovar differentiation and has potential for the rapid, simultaneous identification of both species and serovar of an unknown Listeria isolate by simply measuring a whole-cell infrared spectrum.     

Biosynthesis pathways for deoxysugars in antibiotic-producing actinomycetes: isolation, characterization and generation of novel glycosylated derivatives

Many bioactive natural products synthesized by actinomycetes are glycosylated compounds in which the appended sugars contribute to specific interactions with their biological target. Most of these sugars are 6-deoxyhexoses, of which more than 70 different forms have been identified, and an increasing number of gene clusters involved in 6-deoxyhexoses biosynthesis are being characterized from antibiotic-producing actinomycetes. Novel glycosylated compounds have been generated by modifying natural deoxysugar biosynthesis pathways in the producer organisms, and/or the simultaneous expression in these strains of selected deoxysugar biosynthesis genes from other strains. Non-producing strains endowed with the capacity to synthesize novel deoxysugars through the expression of engineered deoxysugar biosynthesis clusters can also be used as alternative hosts. Transfer of these deoxysugars to a multiplicity of aglycones relies upon the existence of glycosyltransferases with an inherent degree of 'relaxed substrate specificity'. In this review, we analyze how the knowledge coming out from isolation and characterization of deoxysugar biosynthesis pathways from actinomycetes is being used to produce novel glycosylated derivatives of natural products.     

Acoustical detection of coronary occlusions using neural networks

A nonlinear neural network classifier was applied to noninvasive acoustic detection of coronary artery disease; the classifier included a feature vector, derived from diastolic heart sounds, and a multi-layered network trained by the backpropagation. The feature vector is based on the linear prediction coefficients of the autoregressive method after an adaptive line enhancement method was used as the input pattern to the neural network. One hundred and twelve recordings (70 abnormal, 42 normal) were studied and the network was trained on a randomly chosen set of six abnormal and six normal patients. It was tested on a database consisting of 100 recordings to which it had not been exposed. The network correctly identified 50 of the 64 patients with coronary artery disease and 32 of the 36 patients without any coronary artery occlusions. These results showed that this neural network is capable of distinguishing normal patients from abnormal patients. In addition, the diagnostic capability of this approach is much better than any other available noninvasive approach.