Solving unobserved heterogeneity with latent class inflated Poisson regression model.

Inflated data and over-dispersion are two common problems when modeling count data with traditional Poisson regression models. In this study, we propose a latent class inflated Poisson (LCIP) regression model to solve the unobserved heterogeneity that leads to inflations and over-dispersion. The performance of the model estimation is evaluated through simulation studies. We illustrate the usefulness of introducing a latent class variable by analyzing the Behavioral Risk Factor Surveillance System (BRFSS) data, which contain several excessive values and characterized by over-dispersion. As a result, the new model we proposed displays a better fit than the standard Poisson regression and zero-inflated Poisson regression models for the inflated counts.

Modeling health survey data with excessive zero and K responses.

Zero-inflated Poisson regression is a popular tool used to analyze data with excessive zeros. Although much work has already been performed to fit zero-inflated data, most models heavily depend on special features of the individual data. To be specific, this means that there is a sizable group of respondents who endorse the same answers making the data have peaks. In this paper, we propose a new model with the flexibility to model excessive counts other than zero, and the model is a mixture of multinomial logistic and Poisson regression, in which the multinomial logistic component models the occurrence of excessive counts, including zeros, K (where K is a positive integer) and all other values. The Poisson regression component models the counts that are assumed to follow a Poisson distribution. Two examples are provided to illustrate our models when the data have counts containing many ones and sixes. As a result, the zero-inflated and K-inflated models exhibit a better fit than the zero-inflated Poisson and standard Poisson regressions.

An improved method of xylose utilization by recombinant Saccharomyces cerevisiae.

The aim of this study was to develop a method to optimize expression levels of xylose-metabolizing enzymes to improve xylose utilization capacity of Saccharomyces cerevisiae. A xylose-utilizing recombinant S. cerevisiae strain YY2KL, able to express nicotinamide adenine dinucleotide phosphate, reduced (NADPH)-dependent xylose reductase (XR), nicotinamide adenine dinucleotide (NAD(+))-dependent xylitol dehydrogenase (XDH), and xylulokinase (XK), showed a low ethanol yield and sugar consumption rate. To optimize xylose utilization by YY2KL, a recombinant expression plasmid containing the XR gene was transformed and integrated into the aur1 site of YY2KL. Two recombinant expression plasmids containing an nicotinamide adenine dinucleotide phosphate (NADP(+))-dependent XDH mutant and XK genes were dually transformed and integrated into the 5S ribosomal DNA (rDNA) sites of YY2KL. This procedure allowed systematic construction of an S. cerevisiae library with different ratios of genes for xylose-metabolizing enzymes, and well-grown colonies with different xylose fermentation capacities could be further selected in yeast protein extract (YPX) medium (1 % yeast extract, 2 % peptone, and 2 % xylose). We successfully isolated a recombinant strain with a superior xylose fermentation capacity and designated it as strain YY5A. The xylose consumption rate for strain YY5A was estimated to be 2.32 g/gDCW/h (g xylose/g dry cell weight/h), which was 2.34 times higher than that for the parent strain YY2KL (0.99 g/gDCW/h). The ethanol yield was also enhanced 1.83 times by this novel method. Optimal ratio and expression levels of xylose-metabolizing enzymes are important for efficient conversion of xylose to ethanol. This study provides a novel method that allows rapid and effective selection of ratio-optimized xylose-utilizing yeast strains. This method may be applicable to other multienzyme systems in yeast.

Pilot-scale ethanol production from rice straw hydrolysates using xylose-fermenting Pichia stipitis.

Ethanol was produced at pilot scale from rice straw hydrolysates using a Pichia stipitis strain previously adapted to NaOH-neutralized hydrolysates. The highest ethanol yield was 0.44 ± 0.02 g(p)/g(s) at an aeration rate of 0.05 vvm using overliming-detoxified hydrolysates. The yield with hydrolysates conditioned by ammonia and NaOH was 0.39 ± 0.01 and 0.34 ± 0.01 g(p)/g(s), respectively, were achieved at the same aeration rate. The actual ethanol yield from hydrolysate fermentation with ammonia neutralization was similar to that with overliming hydrolysate after taking into account the xylose loss resulting from these conditioning processes. Moreover, the ethanol yield from ammonia-neutralized hydrolysates could be further enhanced by increasing the initial cell density by two-fold or reducing the combined concentration of furfural and 5-hydroxymethyl furfural to 0.6g/L by reducing the severity of operational conditions in pretreatment. This study demonstrated the potential for commercial ethanol production from rice straw via xylose fermentation.

Screening of dengue virus in field-caught Aedes aegypti and Aedes albopictus (Diptera: Culicidae) by one-step SYBR green-based reverse transcriptase-polymerase chain reaction assay during 2004-2007 in Southern Taiwan.

We carried out virological surveillance of dengue virus (DENV) in field-caught Aedes mosquitoes during 2004-2007 to estimate the monthly prevalence of infected females in dengue high-risk areas of Taiwan. A total of 92,892 Aedes aegypti (43,133 females and 49,759 males) and 79,315 Aedes albopictus (57,319 females and 21,996 males) adults were collected, grouped into 25,654 pools, and processed for virus detection using a one-step SYBR Green-based real-time reverse transcriptase-polymerase chain reaction assay. DENVs were periodically and sympatrically detected in Ae. aegypti females in accordance with major dengue outbreaks and the corresponding dengue serotypes. Only 0.2% of 7628 pools of Ae. aegypti females were positive for DENVs. This resulted in an overall estimated infection rate (maximum likelihood estimation) of 0.970 per 1000 mosquitoes (95% confidence interval [CI] = 0.53-1.65). The total monthly infection rates ranged from 0.50 to 2.23 per 1000 mosquitoes (95% CI = 0.03-10.71). When sampling areas were scaled down to the city level, monthly infection rates increased to 0.73-12.59 (95% CI = 0.06-59.19). Monthly infection rates over all sampling areas and at the city level increased significantly by month. All positive pools were collected in July (one pool), August (two pools), September (one pool), October (three pools), November (four pools), and December (one pool). All four virus serotypes were detected in mosquitoes, which were consistent with dengue serotypes infecting humans in 2004 (DENV-4), 2005 and 2006 (DENV-2 and DENV-3), and 2007 (DENV-1). Our results provide supporting evidence that, in general, DENV infection rates were low in local Aedes mosquito population during 2004-2007 and that transovarial transmission may not be occurring or is occurring at much lower rates than evidenced in some endemic countries.

Epidemiology of diabetes mellitus among First Nations and non-First Nations adults.

BACKGROUND: First Nations people in Canada experience a disproportionate burden of type 2 diabetes mellitus. To increase our understanding of this evolving epidemic, we compared the epidemiology of diabetes between First Nations and non-First Nations adults in Saskatchewan from 1980 to 2005. METHODS: We used administrative databases to perform a population-based study of diabetes frequency, incidence and prevalence in adults by ethnic background, year, age and sex. RESULTS: We identified 8275 First Nations and 82,306 non-First Nations people with diabetes from 1980 to 2005. Overall, the incidence and prevalence of diabetes were more than 4 times higher among First Nations women than among non-First Nations women and more than 2.5 times higher among First Nations men than among non-First Nations men. The number of incident cases of diabetes was highest among First Nations people aged 40-49, while the number among non-First Nations people was greatest in those aged 70 or more years. The prevalence of diabetes increased over the study period from 9.5% to 20.3% among First Nations women and from 4.9% to 16.0% among First Nations men. Among non-First Nations people, the prevalence increased from 2.0% to 5.5% among women and from 2.0% to 6.2% among men. By 2005, almost 50% of First Nations women and more than 40% of First Nations men aged 60 or older had diabetes, compared with less than 25% of non-First Nations men and less than 20% of non-First Nations women aged 80 or older. INTERPRETATION: First Nations adults are experiencing a diabetes epidemic that disproportionately affects women during their reproductive years. This ethnicity-based pattern suggests diverse underlying mechanisms that may include differences in the diabetogenic impact of gestational diabetes.

Separation of furans and carboxylic acids from sugars in dilute acid rice straw hydrolyzates by nanofiltration.

This work studied the concentration of hydrolyzates obtained from dilute acid hydrolysis of rice straw using nanofiltration (NF). In order to minimize the Donnan exclusion effect of the membrane, the hydrolyzate solution was controlled at low pH value. Negative retentions of both furans and carboxylic acids were observed. The maximum separation factor of acetic acid over xylose was 49, while the maximum separation factor of acetic acid over arabinose was 52, when the system was operated at pH 2.9 and an applied pressure of 24.5-34.3 bar. The separation factors of inhibitors over glucose became infinity due to the complete retention of glucose. The separation performance decreased when the operating temperature was increased from 25 to 40 degrees C. The flux deterioration was recovered by flushing with 0.01 N of NaOH and water.

Enhanced ethanol production by fermentation of rice straw hydrolysate without detoxification using a newly adapted strain of Pichia stipitis.

An enhanced inhibitor-tolerant strain of Pichia stipitis was successfully developed through adaptation to acid-treated rice straw hydrolysate. The ethanol production obtained by fermentation of NaOH-neutralized hydrolysate without detoxification using the adapted P. stipitis was comparable to fermentation of overliming-detoxified hydrolysate. The ethanol yield using the adapted P. stipitis with both types of hydrolysate at pH 5.0 achieved 0.45 g(p) g(s)(-1), which is equivalent to 87% of the maximum possible ethanol conversion. Furthermore, the newly adapted P. stipitis demonstrated significantly enhanced tolerance to sulfate and furfural despite the fact that both inhibitors had not been removed from the hydrolysate by NaOH neutralization. Finally, the ethanol conversion could be maintained at 60% and above when the neutralized hydrolysate contained 3.0% sulfate and 1.3gL(-1) furfural.

Model evaluation of the phytoextraction potential of heavy metal hyperaccumulators and non-hyperaccumulators.

Evaluation of the remediation ability of zinc/cadmium in hyper- and non-hyperaccumulator plant species through greenhouse studies is limited. To bridge the gap between greenhouse studies and field applications for phytoextraction, we used published data to examine the partitioning of heavy metals between plants and soil (defined as the bioconcentration factor). We compared the remediation ability of the Zn/Cd hyperaccumulators Thlaspi caerulescens and Arabidopsis halleri and the non-hyperaccumulators Nicotiana tabacum and Brassica juncea using a hierarchical linear model (HLM). A recursive algorithm was then used to evaluate how many harvest cycles were required to clean a contaminated site to meet Taiwan Environmental Protection Agency regulations. Despite the high bioconcentration factor of both hyperaccumulators, metal removal was still limited because of the plants' small biomass. Simulation with N. tabacum and the Cadmium model suggests further study and development of plants with high biomass and improved phytoextraction potential for use in environmental cleanup.

Comparative analysis of full genomic sequences among different genotypes of dengue virus type 3.

BACKGROUND: Although the previous study demonstrated the envelope protein of dengue viruses is under purifying selection pressure, little is known about the genetic differences of full-length viral genomes of DENV-3. In our study, complete genomic sequencing of DENV-3 strains collected from different geographical locations and isolation years were determined and the sequence diversity as well as selection pressure sites in the DENV genome other than within the E gene were also analyzed. RESULTS: Using maximum likelihood and Bayesian approaches, our phylogenetic analysis revealed that the Taiwan's indigenous DENV-3 isolated from 1994 and 1998 dengue/DHF epidemics and one 1999 sporadic case were of the three different genotypes - I, II, and III, each associated with DENV-3 circulating in Indonesia, Thailand and Sri Lanka, respectively. Sequence diversity and selection pressure of different genomic regions among DENV-3 different genotypes was further examined to understand the global DENV-3 evolution. The highest nucleotide sequence diversity among the fully sequenced DENV-3 strains was found in the nonstructural protein 2A (mean +/- SD: 5.84 +/- 0.54) and envelope protein gene regions (mean +/- SD: 5.04 +/- 0.32). Further analysis found that positive selection pressure of DENV-3 may occur in the non-structural protein 1 gene region and the positive selection site was detected at position 178 of the NS1 gene. CONCLUSION: Our study confirmed that the envelope protein is under purifying selection pressure although it presented higher sequence diversity. The detection of positive selection pressure in the non-structural protein along genotype II indicated that DENV-3 originated from Southeast Asia needs to monitor the emergence of DENV strains with epidemic potential for better epidemic prevention and vaccine development.

Retrospective serological study on sequential dengue virus serotypes 1 to 4 epidemics in Tainan City, Taiwan, 1994 to 2000.

BACKGROUND AND PURPOSE: We previously reported the development of a non-structural protein NS1 serotype-specific immunoglobulin G (IgG) enzyme-linked immunosorbent assay (ELISA) for dengue serodiagnosis and seroepidemiological study. This assay can be used to differentiate the immunologic status of individuals into naive, primary, or secondary dengue virus (DENV) infection and identify the DENV serotypes of primary infection. A retrospective study was conducted to investigate the serological responses of confirmed dengue cases infected during each of the sequential DENV-1 (August 1994 to February 1995), DENV-2 (August to December 1997), DENV-3 (August 1998 to January 1999), and DENV-4 (June to December 2000) epidemics in Tainan City, Taiwan. METHODS: 218 serum samples collected 1.1 to 7.2 years postinfection were analyzed by NS1 serotype-specific IgG ELISA together with corresponding acute and/or convalescent serum samples when available. The immunological status and the infecting DENV serotypes were determined for these individuals. RESULTS: High titers of dengue NS1 serotype-specific IgG antibody could be detected in serum samples. Differentiation of immunological status showed that 76.6% and 23.4% of cases had primary and secondary infections, respectively. A significant age-dependent increase in the rate of secondary infection was observed for those cases born before 1942. Notably, analysis of postinfection serum samples of 17 dengue hemorrhagic fever patients infected during the 1998 DENV-3 epidemic showed that 9 cases (53%) had primary infections. CONCLUSIONS: Our data revealed that a majority of the population born after 1943 in Tainan City are naive to DENV infection and are at high risk of infection with all 4 DENV serotypes.

Seaport as a source of hantavirus: a study on isolated isles.

Since small mammals from seaports have relative higher seroprevalences of hantavirus, this seroepidemiologic study was conducted on the isles in Kinmen and Lienchiang Counties along the coast of southern mainland China to determine whether seaport may play the role as a source of hantavirus. Among six species of small mammals trapped in Kinmen County, only male Mus musculus (6.7%) and Rattus losea (5.3%) were found to be positive. In Lienchiang County, five species of small mammals were trapped and positive findings were obtained only in male R. norvegicus (9.3%), male M. musculus (7.7%), and female R. losea (1.6%). There was no significant difference in the seroprevalence between the two counties (Kinmen 3.8% vs. Lienchiang 3.9%). The positive rate in Liaolo (17.9%) of Kinmen County was significantly higher than the remaining trapping stations and those in Matsu Distillery (10.3%) and Fushing (5.8%) of Lienchiang County were significantly higher than the remaining ones. Moreover, a significant inverse correlation was found between the seropositive rate and the distance of small mammal sampling sites to the seaport (p < 0.01). These findings suggest the role of seaport as a source of hantavirus.

Missing data imputation in quality-of-life assessment: imputation for WHOQOL-BREF.

INTRODUCTION: This study investigated the effects of imputing missing data in the WHO Quality of Life Abbreviated Questionnaire (WHOQOL-BREF). The imputation results from both the item and domain levels were compared and the impact of the missing data rate and the number of items included for imputation were examined. METHODS: An empirical analysis and a simulation study were used to examine the effects of missing data rates and the number of items used for imputation on the accuracy for imputation. In the empirical analysis, both item-level and domain-level imputations were performed, and the missing values were imputed using different amounts of data. In the simulation study, sets of 2%, 5% and 10% of the data were drawn randomly and replaced with missing values. Twenty datasets were generated for each situation. The data were imputed and the accuracy of the imputation was reported. RESULTS: In the empirical study, the number of items used for imputation had only a small impact on the accuracy of imputation. Furthermore, in the simulation study, the accuracy rates of imputation did not significantly change as the proportions of missing data increased. However, the number of items used in the computation did contribute to some extent to the missing values imputed. Extreme responses had the worst computations and the lowest accuracy rates. CONCLUSION: It is recommended that as many items as possible be included for imputation within the same domain. However, it is not particularly helpful to use items from different domains for imputation. Researchers should exercise extra caution in interpreting the imputed values of extreme responses.

A retrospective study of hantavirus infection in Kinmen, Taiwan.

In this retrospective study, 140 serum samples collected from 85 scrub typhus-negative patients in Kinmen Island in 2000 were tested for antibodies to hantavirus using enzyme-linked immunosorbent assay. Seven patients (8.23%) were confirmed as having hantavirus infection as demonstrated by increased hantavirus-specific immunoglobulin M and/or immunoglobulin G antibodies in their convalescent serum samples. Analysis of indirect immunofluorescence assay showed that Seoul type was the etiologic agent. Serosurvey of rodents caught in the resident township of these hantavirus-infected human cases showed that the seroprevalence of antibodies to hantavirus among Rattus norvegicus, Mus musculus, and R. flavipectus was 50% (4/8), 20% (1/5), and 2% (7/348), respectively. Molecular analysis showed that these reservoir hosts carried a Seoul type hantavirus. To our knowledge, this is the first report demonstrating indigenous hantavirus cases in Kinmen.

Fever screening at airports and imported dengue.

Airport fever screening in Taiwan, July 2003-June 2004, identified 40 confirmed dengue cases. Results obtained by capture immunoglobulin (Ig) M and IgG enzyme-linked immunoassay, real time 1-step polymerase chain reaction, and virus isolation showed that 33 (82.5%) of 40 patients were viremic. Airport fever screening can thus quickly identify imported dengue cases.

SARS in Taiwan: an overview and lessons learned.

OBJECTIVES: This report aims to describe the epidemiology of severe acute respiratory syndrome (SARS) in Taiwan between March and July 2003, and to examine the public health response. METHODS: Surveillance for SARS was initiated on 14 March 2003. Response activities are described for the isolation of patients; contact tracing; quarantine of contact persons; fever screening for inbound and outbound passengers at the airport; and hospital infection control as assessed by mobile SARS containment teams. RESULTS: Between 14 March and 30 July 2003 a total of 668 probable cases of SARS were reported. Of the 668 cases, 181 (27%) were fatal. Compared to the survivors, fatal cases were more likely to be older (p < 0.001), male (p < 0.05), exposed through hospital contact (p < 0.001), and have a coexisting medical disorder (p<0.001). Between 28 March and 30 July a total of 151,270 persons were quarantined. Among them, 46 (3.0/10,000) were subsequently classified as being probable SARS cases. At the time of the mobile team assessments, 46 (53%) hospitals had implemented WHO infection control recommendations. CONCLUSIONS: In this outbreak, an emergency plan consisted of patient isolation and strict hospital infection control.

Childhood severe acute respiratory syndrome in Taiwan and how to differentiate it from childhood influenza infection.

OBJECTIVE: To investigate clinical features and outcomes of children in Taiwan with laboratory-confirmed severe acute respiratory syndrome (SARS) vs those of children with influenza to differentiate the 2 diseases. DESIGN, SETTING, AND PARTICIPANTS: Patients 20 years or younger with clinical, epidemiological, and laboratory evidence of SARS from March to July 2003 vs children with virus culture-confirmed influenza in a 1:1 age- and sex-matched control group. MAIN OUTCOME MEASURES: Rates of symptoms, abnormal laboratory data, and outcomes of recovery, sequelae, or death. RESULTS: The 15 SARS patients (9 girls and 6 boys) had a median age of 17 years (age range, 4-20 years). Nine patients (60%) were infected through household contact, 4 (27%) nosocomially, 1 (7%) through contact with a neighbor, and 1 (7%) after returning from Hong Kong. All 15 patients had fever, 3 (20%) had chills, and 11 (73%) had cough. Only 1 patient (7%) had sputum production; 1 (7%) had rhinorrhea. At presentation, 5 patients (33%) had leukopenia, 6 (40%) had lymphopenia, and 5 (33%) had monocytopenia. All children recovered without sequelae. Children with SARS had significantly lower incidences of rhinorrhea (odds ratio [OR], 0.01; 95% confidence interval [CI], 0.00-0.09), sputum production (OR, 0.10; 95% CI, 0.02-0.63), and sore throat (OR, 0.17; 95% CI, 0.03-0.85) than children with influenza. Both groups had similar incidences of leukopenia or lymphopenia, but SARS patients had a significantly higher incidence of monocytopenia (33% vs 0%, P = .04). CONCLUSIONS: Childhood SARS is usually not fatal. The absence of rhinorrhea and presence of monocytopenia in SARS may distinguish it from influenza.