Unmixing hyperspectral data: independent and dependent component analysis

The development of high spatial resolution airborne and spaceborne sensors has improved the capability of ground-based data collection in the fields of agriculture, geography, geology, mineral identification, detection [2, 3], and classification [4–8]. The signal read by the sensor from a given spatial element of resolution and at a given spectral band is a mixing of components originated by the constituent substances, termed endmembers, located at that element of resolution. This chapter addresses hyperspectral unmixing, which is the decomposition of the pixel spectra into a collection of constituent spectra, or spectral signatures, and their corresponding fractional abundances indicating the proportion of each endmember present in the pixel [9, 10]. Depending on the mixing scales at each pixel, the observed mixture is either linear or nonlinear [11, 12]. The linear mixing model holds when the mixing scale is macroscopic [13]. The nonlinear model holds when the mixing scale is microscopic (i.e., intimate mixtures) [14, 15]. The linear model assumes negligible interaction among distinct endmembers [16, 17]. The nonlinear model assumes that incident solar radiation is scattered by the scene through multiple bounces involving several endmembers [18]. Under the linear mixing model and assuming that the number of endmembers and their spectral signatures are known, hyperspectral unmixing is a linear problem, which can be addressed, for example, under the maximum likelihood setup [19], the constrained least-squares approach [20], the spectral signature matching [21], the spectral angle mapper [22], and the subspace projection methods [20, 23, 24]. Orthogonal subspace projection [23] reduces the data dimensionality, suppresses undesired spectral signatures, and detects the presence of a spectral signature of interest. The basic concept is to project each pixel onto a subspace that is orthogonal to the undesired signatures. As shown in Settle [19], the orthogonal subspace projection technique is equivalent to the maximum likelihood estimator. This projection technique was extended by three unconstrained least-squares approaches [24] (signature space orthogonal projection, oblique subspace projection, target signature space orthogonal projection). Other works using maximum a posteriori probability (MAP) framework [25] and projection pursuit [26, 27] have also been applied to hyperspectral data. In most cases the number of endmembers and their signatures are not known. Independent component analysis (ICA) is an unsupervised source separation process that has been applied with success to blind source separation, to feature extraction, and to unsupervised recognition [28, 29]. ICA consists in finding a linear decomposition of observed data yielding statistically independent components. Given that hyperspectral data are, in given circumstances, linear mixtures, ICA comes to mind as a possible tool to unmix this class of data. In fact, the application of ICA to hyperspectral data has been proposed in reference 30, where endmember signatures are treated as sources and the mixing matrix is composed by the abundance fractions, and in references 9, 25, and 31–38, where sources are the abundance fractions of each endmember. In the first approach, we face two problems: (1) The number of samples are limited to the number of channels and (2) the process of pixel selection, playing the role of mixed sources, is not straightforward. In the second approach, ICA is based on the assumption of mutually independent sources, which is not the case of hyperspectral data, since the sum of the abundance fractions is constant, implying dependence among abundances. This dependence compromises ICA applicability to hyperspectral images. In addition, hyperspectral data are immersed in noise, which degrades the ICA performance. IFA [39] was introduced as a method for recovering independent hidden sources from their observed noisy mixtures. IFA implements two steps. First, source densities and noise covariance are estimated from the observed data by maximum likelihood. Second, sources are reconstructed by an optimal nonlinear estimator. Although IFA is a well-suited technique to unmix independent sources under noisy observations, the dependence among abundance fractions in hyperspectral imagery compromises, as in the ICA case, the IFA performance. Considering the linear mixing model, hyperspectral observations are in a simplex whose vertices correspond to the endmembers. Several approaches [40–43] have exploited this geometric feature of hyperspectral mixtures [42]. Minimum volume transform (MVT) algorithm [43] determines the simplex of minimum volume containing the data. The MVT-type approaches are complex from the computational point of view. Usually, these algorithms first find the convex hull defined by the observed data and then fit a minimum volume simplex to it. Aiming at a lower computational complexity, some algorithms such as the vertex component analysis (VCA) [44], the pixel purity index (PPI) [42], and the N-FINDR [45] still find the minimum volume simplex containing the data cloud, but they assume the presence in the data of at least one pure pixel of each endmember. This is a strong requisite that may not hold in some data sets. In any case, these algorithms find the set of most pure pixels in the data. Hyperspectral sensors collects spatial images over many narrow contiguous bands, yielding large amounts of data. For this reason, very often, the processing of hyperspectral data, included unmixing, is preceded by a dimensionality reduction step to reduce computational complexity and to improve the signal-to-noise ratio (SNR). Principal component analysis (PCA) [46], maximum noise fraction (MNF) [47], and singular value decomposition (SVD) [48] are three well-known projection techniques widely used in remote sensing in general and in unmixing in particular. The newly introduced method [49] exploits the structure of hyperspectral mixtures, namely the fact that spectral vectors are nonnegative. The computational complexity associated with these techniques is an obstacle to real-time implementations. To overcome this problem, band selection [50] and non-statistical [51] algorithms have been introduced. This chapter addresses hyperspectral data source dependence and its impact on ICA and IFA performances. The study consider simulated and real data and is based on mutual information minimization. Hyperspectral observations are described by a generative model. This model takes into account the degradation mechanisms normally found in hyperspectral applications—namely, signature variability [52–54], abundance constraints, topography modulation, and system noise. The computation of mutual information is based on fitting mixtures of Gaussians (MOG) to data. The MOG parameters (number of components, means, covariances, and weights) are inferred using the minimum description length (MDL) based algorithm [55]. We study the behavior of the mutual information as a function of the unmixing matrix. The conclusion is that the unmixing matrix minimizing the mutual information might be very far from the true one. Nevertheless, some abundance fractions might be well separated, mainly in the presence of strong signature variability, a large number of endmembers, and high SNR. We end this chapter by sketching a new methodology to blindly unmix hyperspectral data, where abundance fractions are modeled as a mixture of Dirichlet sources. This model enforces positivity and constant sum sources (full additivity) constraints. The mixing matrix is inferred by an expectation-maximization (EM)-type algorithm. This approach is in the vein of references 39 and 56, replacing independent sources represented by MOG with mixture of Dirichlet sources. Compared with the geometric-based approaches, the advantage of this model is that there is no need to have pure pixels in the observations. The chapter is organized as follows. Section 6.2 presents a spectral radiance model and formulates the spectral unmixing as a linear problem accounting for abundance constraints, signature variability, topography modulation, and system noise. Section 6.3 presents a brief resume of ICA and IFA algorithms. Section 6.4 illustrates the performance of IFA and of some well-known ICA algorithms with experimental data. Section 6.5 studies the ICA and IFA limitations in unmixing hyperspectral data. Section 6.6 presents results of ICA based on real data. Section 6.7 describes the new blind unmixing scheme and some illustrative examples. Section 6.8 concludes with some remarks.

Hyperspectral unmixing algorithm via dependent component analysis

This paper introduces a new method to blindly unmix hyperspectral data, termed dependent component analysis (DECA). This method decomposes a hyperspectral images into a collection of reflectance (or radiance) spectra of the materials present in the scene (endmember signatures) and the corresponding abundance fractions at each pixel. DECA assumes that each pixel is a linear mixture of the endmembers signatures weighted by the correspondent abundance fractions. These abudances are modeled as mixtures of Dirichlet densities, thus enforcing the constraints on abundance fractions imposed by the acquisition process, namely non-negativity and constant sum. The mixing matrix is inferred by a generalized expectation-maximization (GEM) type algorithm. This method overcomes the limitations of unmixing methods based on Independent Component Analysis (ICA) and on geometrical based approaches. The effectiveness of the proposed method is illustrated using simulated data based on U.S.G.S. laboratory spectra and real hyperspectral data collected by the AVIRIS sensor over Cuprite, Nevada.

Lysotypes and plasmidial profile of Salmonella Serovar Typhimurium isolated from children with enteric processes in the cities of Rio de Janeiro, RJ, and Salvador, BA - Brazil

The lysotypes, plasmidial profiles, and profiles of resistance to antimicrobial agents were determined in 111 Salmonella Typhimurium strains isolated from feces and blood of children treated in Rio de Janeiro and in Salvador. Six distinct lysotypes (19, 41, 97, 105, 120 and 193) were recognized, with a predominance of lysotype 193 (59.7%) in Rio de Janeiro and of phage type 105 (38.4) in Salvador. Approximately 86.7% of the lysotype 193 strains presented multiple resistance to more than six antimicrobial agents, whereas 93% of lysotype 105 strains were fully susceptible. More than 90% of the strains presented plasmids distributed into 36 different profiles in Rio de Janeiro and into 10 profiles in Salvador. A 40 MDa plasmid was the most frequent (47%) in the strains from Rio de Janeiro, whereas a 61 MDa plasmid predominated (14.5%) in Salvador. Combined analysis of plasmid profile and classification into lysotypes (especially those belonging to types 105 and 103, proved to be more discriminatory than the other methods applied).

Effect of skin colour and selected physical characteristics on Schistosoma mansoni dependent morbidity

The effect of the colour group on the morbidity due to Schistosoma mansoni was examined in two endemic areas situated in the State of Minas Gerais, Brazil. Of the 2773 eligible inhabitants, 1971 (71.1%) participated in the study: 545 (27.6%) were classified as white, 719 (36.5%) as intermediate and 707 (35.9%) as black. For each colour group, signs and symptoms of individuals who eliminated S.mansoni eggs (cases) were compared to those who did not present eggs in the faeces (controls). The odds ratios were adjusted by age, gender, previous treatment for schistosomiasis, endemic area and quality of the household. There was no evidence of a modifier effect of colour on diarrhea, bloody faeces or abdominal pain. A modifier effect of colour on hepatomegaly was evident among those heaviest infected (> 400 epg): the adjusted odds ratios for palpable liver at the middle clavicular and the middle sternal lines were smaller among blacks (5.4 and 6.5, respectively) and higher among whites (10.6 and 12.9) and intermediates (10.4 and 10.1, respectively). These results point out the existence of some degree of protection against hepatomegaly among blacks heaviest infected in the studied areas.

Progression of Salmonella Enteritidis phage type 4 strains in São Paulo State, Brazil

A total of 574 S. Enteritidis strains (383 from human sources and 191 from non-human sources) isolated between 1975-95, in São Paulo State, Brazil, were phagetyped. Among the strains isolated during the period of 1975-92, 80.9% of them belonged to phage type 8 (PT-8), but in 1993 strains of PT-4 accounted for 65.2% of all the S. Enteritidis isolates. In the following years, PT-4 strains accounted for 99.7% and 98.4% of phagetyped S. Enteritidis strains. The results obtained suggested that the current epidemic of S. Enteritidis in São Paulo State is clearly associated with the progression of PT-4 strains.

Changing patterns of Salmonella serovars: increase of Salmonella Enteritidis in São Paulo, Brazil

Serovars of a total of 5,490 Salmonella strains isolated during the period of 1991-95, from human infections (2,254 strains) and from non-human materials (3,236 strains) were evaluated. In the studied period, 81 different serovars were determined among human isolates. Salmonella Enteritidis corresponded to 1.2% in 1991, 2% in 1992, 10.1% in 1993, 43.3% in 1994, and 64.9% in 1995 of all isolates. A significant rise on the isolation of this serovar was seen since 1993 linked to food poisoning outbreaks. It is reported also an increase on the isolation of S. Enteritidis from blood cultures, associated mainly with patients with immunodeficiency syndrome. S. Enteritidis was prevalent among one hundred and thirty different serovars isolated from non-human sources. Increasing number of isolation of this serovar was seen from shell eggs, breeding flocks and from environmental samples. It is also reported a contamination of commercial feed stuffs by S. Enteritidis which represents a major concern for Brazilian poultry industry.

Lettuce (Lactuca sativa L.) leaf-proteome profiles after exposure to cylindrospermopsin and a microcystin-LR/cylindrospermopsin mixture: A concentration-dependent response

The intensification of agricultural productivity is an important challenge worldwide. However, environmental stressors can provide challenges to this intensification. The progressive occurrence of the cyanotoxins cylindrospermopsin (CYN) and microcystin-LR (MC-LR) as a potential consequence of eutrophication and climate change is of increasing concern in the agricultural sector because it has been reported that these cyanotoxins exert harmful effects in crop plants. A proteomic-based approach has been shown to be a suitable tool for the detection and identification of the primary responses of organisms exposed to cyanotoxins. The aim of this study was to compare the leaf-proteome profiles of lettuce plants exposed to environmentally relevant concentrations of CYN and a MC-LR/CYN mixture. Lettuce plants were exposed to 1, 10, and 100 lg/l CYN and a MC-LR/CYN mixture for five days. The proteins of lettuce leaves were separated by twodimensional electrophoresis (2-DE), and those that were differentially abundant were then identified by matrix-assisted laser desorption/ionization time of flight-mass spectrometry (MALDI-TOF/TOF MS). The biological functions of the proteins that were most represented in both experiments were photosynthesis and carbon metabolism and stress/defense response. Proteins involved in protein synthesis and signal transduction were also highly observed in the MC-LR/CYN experiment. Although distinct protein abundance patterns were observed in both experiments, the effects appear to be concentration-dependent, and the effects of the mixture were clearly stronger than those of CYN alone. The obtained results highlight the putative tolerance of lettuce to CYN at concentrations up to 100 lg/l. Furthermore, the combination of CYN with MC-LR at low concentrations (1 lg/l) stimulated a significant increase in the fresh weight (fr. wt) of lettuce leaves and at the proteomic level resulted in the increase in abundance of a high number of proteins. In contrast, many proteins exhibited a decrease in abundance or were absent in the gels of the simultaneous exposure to 10 and 100 lg/l MC-LR/CYN. In the latter, also a significant decrease in the fr. wt of lettuce leaves was obtained. These findings provide important insights into the molecular mechanisms of the lettuce response to CYN and MC-LR/CYN and may contribute to the identification of potential protein markers of exposure and proteins that may confer tolerance to CYN and MC-LR/CYN. Furthermore, because lettuce is an important crop worldwide, this study may improve our understanding of the potential impact of these cyanotoxins on its quality traits (e.g., presence of allergenic proteins).

Salmonella-S. mansoni ASSOCIATION IN PATIENTS WITH ACQUIRED IMMUNODEFICIENCY SYNDROME

Two young men with Salmonella bacteraemia, active schistosomiasis and the acquired immunodeficiency syndrome are reported. The clinical presentation comprised nonspecific signs and symptoms, such as fatigue, malaise, weight loss, diarrhoea, prolonged fever, and hepatosplenomegaly. In one patient, liver biopsy showed poorly formed granulomata around Schistosoma mansoni eggs and hepatitis. Treatment of schistosomiasis alone induced consistent clinical improvement with eventual cure of both Salmonella and S. mansoni infections. Recognition of the Salmonella-S. mansoni association in patients with AIDS is important because treatment of schistosomiasis makes a difference, improving the prognosis of this otherwise, recurrent, potentially fatal bacteraemia.

Culture-dependent characterization of cyanobacterial diversity in the intertidal zones of the Portuguese coast: A polyphasic study

Cyanobacteria are important primary producers, and many are able to fix atmospheric nitrogen playing a key role in the marine environment. However, not much is known about the diversity of cyanobacteria in Portuguese marine waters. This paper describes the diversity of 60 strains isolated from benthic habitats in 9 sites (intertidal zones) on the Portuguese South and West coasts. The strains were characterized by a morphological study (light and electron microscopy) and by a molecular characterization (partial 16S rRNA, nifH, nifK, mcyA, mcyE/ndaF, sxtI genes). The morphological analyses revealed 35 morphotypes (15 genera and 16 species) belonging to 4 cyanobacterial Orders/Subsections. The dominant groups among the isolates were the Oscillatoriales. There is a broad congruence between morphological and molecular assignments. The 16S rRNA gene sequences of 9 strains have less than 97% similarity compared to the sequences in the databases, revealing novel cyanobacterial diversity. Phylogenetic analysis, based on partial 16S rRNA gene sequences showed at least 12 clusters. One-third of the isolates are potential N2-fixers, as they exhibit heterocysts or the presence of nif genes was demonstrated by PCR. Additionally, no conventional freshwater toxins genes were detected by PCR screening.

Desenvolvimento de um método potenciométrico para deteção de salmonella typhimurium

A Salmonella é um microrganismo responsável por grande parte das doenças alimentares, podendo por em causa a saúde pública da área contaminada. Uma deteção rápida, eficiente e altamente sensível e extremamente importante, sendo um campo em franco desenvolvimento e alvo de variados e múltiplos estudos na comunidade cientifica atual. Foi desenvolvido um método potenciométrico para a deteção de Salmonellas, com elétrodos seletivos de iões, construídos em laboratório com pontas de micropipetas, fios de prata e sensores com composição otimizada. O elétrodo indicador escolhido foi um ESI seletivo a cadmio, para redução da probabilidade de interferências no método, devido a pouca abundancia do cadmio em amostras alimentares. Elétrodos seletivos a sódio, elétrodos de Ag/AgCl de simples e de dupla juncão foram também construídos e caracterizados para serem aplicados como elétrodos de referência. Adicionalmente otimizaram-se as condições operacionais para a analise potenciométrica, nomeadamente o elétrodo de referencia utilizado, condicionamento dos elétrodos, efeito do pH e volume da solução amostra. A capacidade de realizar leituras em volumes muito pequenos com limites de deteção na ordem dos micromolares por parte dos ESI de membrana polimérica, foi integrada num ensaio com um formato nao competitivo ELISA tipo sanduiche, utilizando um anticorpo primário ligado a nanopartículas de Fe@Au, permitindo a separação dos complexos anticorpo-antigénio formados dos restantes componentes em cada etapa do ensaio, pela simples aplicação de um campo magnético. O anticorpo secundário foi marcado com nanocristais de CdS, que são bastante estáveis e é fácil a transformação em Cd2+ livre, permitindo a leitura potenciométrica. Foram testadas várias concentrações de peroxido de hidrogénio e o efeito da luz para otimizar a dissolução de CdS. O método desenvolvido permitiu traçar curvas de calibração com soluções de Salmonellas incubadas em PBS (pH 4,4) em que o limite de deteção foi de 1100 CFU/mL e de 20 CFU/mL, utilizando volumes de amostra de 10 ƒÊL e 100 ƒÊL, respetivamente para o intervalo de linearidade de 10 a 108 CFU/mL. O método foi aplicado a uma amostra de leite bovino. A taxa de recuperação media obtida foi de 93,7% } 2,8 (media } desvio padrão), tendo em conta dois ensaios de recuperação efetuados (com duas replicas cada), utilizando um volume de amostra de 100 ƒÊL e concentrações de 100 e 1000 CFU/mL de Salmonella incubada.

Polymerase chain reaction (PCR) for the detection of Salmonella in artificially inoculated chicken meat

The aim of this study was to develop a polymerase chain reaction (PCR) protocol for the detection of Salmonella in artificially contaminated chicken meat. Tests were performed with different dilutions of Salmonella Typhimurium or Salmonella Enteritidis cells (10-7, 10-8 or 10-9 CFU/mL) inoculated in chicken meat samples, in order to establish the limits of detection, incubation times (0, 6, 8 and 24 hours of pre-enrichment in PBW 1%) and three DNA extraction protocols (phenol-chloroform, thermal treatment and thermal treatment and Sephaglass). The assay was able to detect until 10-9 CFU/mL of initial dilution of Salmonella cells inoculated in chicken meat, which allows detection of Salmonella within 48 hours, including 24 hours of pre-enrichment and using the phenol-chloroform DNA extraction protocol. As the results are obtained in a shorter time period than that of microbiological culture, this procedure will be useful in the methodology for detection of Salmonella in chicken.

Isolation of Salmonella enterica and serologic reactivity to Leptospira interrogans in opossums (Didelphis virginiana) from Yucatán, México

The presence of Salmonella enterica and serologic evidence of infection by Leptospira interrogans, were detected in the opossum Didelphis virginiana in a semi-urban locality of the Yucatán State, México. Ninety-one opossums were captured during the period April 1996 and May 1998. From a total of 17 feces samples, four Salmonella enterica subsp. enterica serotypes (Sandiego, Newport, Anatum, and Minnesota), and one Salmonella enterica subsp. arizonae serovar O44:Z4,Z23:- were isolated. Some opossums presented mixed infections. From 81 sera samples, four (4.9%) were positive to antibodies to Leptospira serovars pomona and wolfii. Both animals infected with Salmonella enterica and those serologically positive to Leptospira interrogans were captured in peridomestic habitat. Opossums infected with Salmonella enterica, were captured in dry season, and those seropositive to Leptospira interrogans during the rainy season. The implications of infection and reactivity of these zoonotic pathogens in D. virginiana in the Yucatan state are briefly discussed.

Crystallographic studies on molybdopterin-dependent enzymes

Dissertação para obtenção do Grau de Doutor em Bioquímica, Especialidade Bioquímica Estrutural

Conventional and molecular typing of Salmonella typhi strains from Brazil

Phenotypic and genotypic characteristics of Salmonella Typhi were studied in 30 strains, isolated in different years, from some areas in Brazil. Conventional typing methods were performed by biochemical tests, Vi phage-typing scheme, and antimicrobial susceptibility test. Molecular typing methods were performed by analysis of plasmid DNA and by random amplified polymorphic DNA (RAPD-PCR). For the latter, an optimization step was performed to ensure the reproducibility of the process in genetic characterization of S. Typhi. The predominance of 76.7% of biotype I (xylose +, arabinose -) was noticed in all studied areas. Three phage types were recognized, with prominence for the phage types A (73.3%) and I+IV (23.3%). All the strains were susceptible to the drugs used. However, 36.7% of the strains contained plasmids, with predominance of the 105 Kb plasmid. RAPD was capable of grouping the strains in 8 genotypic patterns using primer 784, in 6, using primer 787 and in 7, using primer 797. Conventional phenotypic typing methods, as well as the DNA plasmid analysis, presented nonsignificant discriminatory power; however, RAPD-PCR analysis showed discriminatory power, reproducibility, easy interpretation and performance, being considered as a promising alternative typing method for S. Typhi.

Phage types of Salmonella Enteritidis isolated from clinical and food samples, and from broiler carcasses in Southern Brazil

272 isolates of Salmonella Enteritidis (111 isolated from frozen broiler chicken carcasses, 126 from human food and other biological materials involved in food poisoning outbreaks and 35 from different poultry materials) were selected for phage typing. From these, 111 were phage typed, 57.65% being classified as phage type 4, 32.43% as phage type 4a, 3.60% as phage type 6a and 0.90% as phage type 7, whereas 5.40% samples were not phage typeable. The predominance of phage type 4 is in agreement with the results published worldwide, and reinforces the need for studies related to the epidemiological meaning of these findings.