Handling missingness when modeling the force of infection from clustered seroprevalence data

Modeling infectious diseases data is a relatively young research area in which clustering and stratification are key features. It is not unlikely for these data to have missing values. If values are missing completely at random, the analysis on the complete cases is valid. However, in practice this assumption is usually not fulfilled. This article shows the effect of ignoring missing data in modeling the force of infection of the bovine herpesvirus-1 in Belgian cattle and proposes the use of weighted generalized estimating equations with constrained fractional polynomials as a flexible modeling tool.     

In situ evaluation of fruit yield and estimation of repeatability coefficient for major fruit traits of umbu tree [Spondias tuberosa (Anacardiaceae)] in the semi-arid region of Brazil

The umbu tree (Spondias tuberosa) is a xerophytic woody species endemic to the northeast of Brazil. The present work reports a three-year in situ evaluation of major yield components in a random set of trees in Petrolina city, Pernambuco State, Brazil and the application of repeatability coefficient in the context of genetic resources utilization. Five traits were assessed: total fruit yield per plant (FY); total number of fruits per plant (NF); average fruit weight (FW); average pulp weight (PW) and average fruit diameter (FD). The values observed for FY ranged from 4.2–184 kg with mean of 61.5 kg. The values for NF ranged from 257 to 12,981 fruits/tree with mean 3,993 fruits. For FW, the range was from 10.82–23.36 g with mean of 16.03 g. The values for PW ranged from 7 to 17 g with mean of 11.2 g, while FD values varied from 2.5 to 3.5 cm with mean of 3.0 cm. Six methods were employed to estimate repeatability coefficient (r) for all parameters. The r ranged from 0.68–0.76 for NF; 0.87–0.89 for FW; 0.65–0.75 for FY; 0.64–0.78 for PW and 0.70–0.84 for FD. Narrower ranges across methods were observed to r-values greater than 0.80, as observed for FW and FD. Three to four years of measurements will be necessary for FY, NF, and PW and one year for FW to obtain a precision of 90% to provide a reliable identification of individuals for in situ or ex situ conservation.     

Incorporating parameter uncertainty into prediction intervals for spatial data modeled via a parametric variogram

In spatial predictions, researchers usually treat the estimated theoretical variogram parameters as known without error and ignore the variability of the parameter estimators. Although the prediction is still unbiased, the prediction error is usually underestimated. Therefore, the coverage probability of the prediction interval usually is lower than the nominal probability. A simulation study is performed to show how the coverage probability for prediction relates to the true range and sill of an exponential variogram. This article proposes two parametric bootstrap methods to incorporate the variability of the corresponding parameter estimators. A simulation study is performed to evaluate the coverage probability of these proposed methods. Finally, we apply the parametric bootstrap methods to a real dataset and compare the results with those from naive (i.e., treating estimated parameters as known) and Bayesian methods.     

Nitrogen fixation in biological soil crusts from southeast Utah,USA

Biological soil crusts can be the dominant source of N for arid land ecosystems. We measured potential N fixation rates biweekly for 2 years, using three types of soil crusts: (1) crusts whose directly counted cells were >98% Microcoleus vaginatus (light crusts); (2) crusts dominated by M. vaginatus, but with 20% or more of the directly counted cells represented by Nostoc commune and Scytonema myochrous (dark crusts); and (3) the soil lichen Collema sp. At all observation times, Collema had higher nitrogenase activity (NA) than dark crusts, which had higher NA than light crusts, indicating that species composition is critical when estimating N inputs. In addition, all three types of crusts generally responded in a similar fashion to climate conditions. Without precipitation within a week of collection, no NA was recorded, regardless of other conditions being favorable. Low (<1°C) and high (>26°C) temperatures precluded NA, even if soils were moist. If rain or snow melt had occurred 3 or less days before collection, NA levels were highly correlated with daily average temperatures of the previous 3 days (r²=0.93 for Collema crusts; r²=0.86 for dark crusts and r²=0.83 for light crusts) for temperatures between 1°C and 26°C. If a precipitation event followed a long dry period, NA levels were lower than if collection followed a time when soils were wet for extended periods (e.g., winter). Using a combination of data from a recording weather datalogger, time-domain reflectometry, manual dry-down curves, and N fixation rates at different temperatures, annual N input from the different crust types was estimated. Annual N input from dark crusts found at relatively undisturbed sites was estimated at 9 kg ha^-1 year^-1. With 20% cover of the N-fixing soil lichen Collema, inputs are estimated at 13 kg ha^-1 year^-1. N input from light crusts, generally indicating soil surface disturbance, was estimated at 1.4 kg ha^-1 year^-1. The rates in light crusts are expected to be highly variable, as disturbance history will determine cyanobacterial biomass and therefore N fixation rates.     

QTL mapping of sheath blight resistance in a deep-water rice cultivar

Sheath blight, caused by the pathogen Rhizoctonia solani Kühn, is one of the most serious diseases of rice and leads to severe yield loss worldwide. A recombinant inbred line (RIL) population consisting of 121 lines was constructed from a cross between HH1B and RSB03, the latter of which is a deep-water rice variety. Five traits were used to evaluate sheath blight resistance, namely disease rating (DR), lesion length (LL), lesion height (LH), relative lesion length [RLL, the ratio of LL to plant height (PH)], and relative LH (RLH, the ratio of LH to PH). Using the RIL population and 123 molecular markers, we identified 28 quantitative trait loci (QTLs) for the five traits in two environments. These QTLs are located on nine chromosomes and most of them are environment specific. A major QTL for DR (qSBR1) on chromosome 1 was identified with contributions of 12.7% at Shanghai and 42.6% at Hainan, and it collocated with a QTL for PH. The allele at this locus from RSB03 enhances sheath blight resistance and increases PH. Another QTL for DR on chromosome 7 was adjacent to QTLs for heading date (HD) and four other disease traits. RSB03 also carries the resistant allele at this locus and shortens HD. The susceptible parent, HH1B, provides the resistance allele at the locus qSBR8, where QTLs for four other disease traits were identified. QTL mapping results showed that most QTLs for LL, LH, RLL, and RLH are collocated with QTLs for DR. Three QTLs for DR are independent from HD, PH, and four other disease traits, while four QTLs are closely related to HD and PH. Four QTLs for LL, LH, RLL, and RLH are independent from DR, HD, and PH, while there is only one region harboring QTLs for these four traits and HD. Correlation analysis and QTL mapping results indicated that LL, LH, RLL, and RLH might be important indices, like DR, for evaluating the level of resistance to rice sheath blight.     

Allele mining and haplotype discovery in barley candidate genes for drought tolerance

In the present study, allele mining was conducted on a panel of drought related candidate genes in a set of 96 barley genotypes using EcoTILLING, which is a variant of the targeting induced local lesions in genomes (TILLING) technology. Analyzing approximately 1.5 million basepairs in barley a total number of 94 verified unique haplotypes were identified in 18 amplicons designed for 9 genes. Overall, 185 single nucleotide polymorphisms (SNPs) and 46 insertions/deletions (INDELs) were detected with a mean of 1SNP/92 bp and 1INDEL/372 bp genomic sequence. Based on overlapping haplotype sequences, markers were developed for four candidate genes (HvARH1, HvSRG6, HvDRF1, HVA1), which allows distinguishing between the main haplotypes showing either differences in amino acid sequence or which have larger INDELs in the promoter region. As “proof of concept”, the HvARH1 and HvSRG6 haplotypes were tested for the level of abscisic acid-induced gene expression in subsets of genotypes belonging to different haplotype categories. An integrated database was developed to contain information about the genes, genotypes, and haplotypes analyzed in this study. The database supplies profound information about the natural variation in the tested drought related candidate genes providing a significant asset for further mapping studies dealing with this highly polygenic trait.     

Impact of epistasis and QTL × environmental interaction on the oil filling rate of soybean seed at different developmental stages

The oil accumulation in the developing soybean seed has been shown to be a dynamic process with different rates and activities at different phases affected by both genotype and environment. The objective of the present study was to investigate additive, epistatic and quantitative trait loci (QTL) × environment interaction (QE) effects of the QTL controlling oil filling rate in soybean seed. A total of 143 recombinant inbred lines (RILs) derived from the cross of Charleston and Dongnong 594 were used in this study to obtain 2 years of field data (2004 and 2005). A total of 26 QTL with significantly unconditional and conditional additive (a) effect and/or additive × environment interaction (ae) effect at different filling stages were identified on 14 linkage groups. Among the QTL with significant a effects, 18 QTL showed positive effects and 6 QTL had negative effects on seed filling rate of oil content during seed development. A total of 29 epistatic pairwise QTL underlying seed filling rate were identified at different filling stages. About 28 pairs of the QTL showed additive × additive epistatic (aa) effects and 14 pairs of the QTL showed aa × environment interaction (aae) effects at different filling stages. QTL with aa and aae (additive × additive × environment) effects appeared to vary at different filling stages. Our results demonstrated that oil filling rate in soybean seed were under genetic, developmental and environmental control.     

Molecular genetics of race non-specific rust resistance in wheat

Over 150 resistance genes that confer resistance to either leaf rust, stripe rust or stem rust have been catalogued in wheat or introgressed into wheat from related species. A few of these genes from the ‘slow-rusting’ adult plant resistance (APR) class confer partial resistance in a race non-specific manner to one or multiple rust diseases. The recent cloning of two of these genes, Lr34/Yr18, a dual APR for leaf rust and stripe rust, and Yr36, a stripe rust APR gene, showed that they differ from other classes of plant resistance genes. Currently, seven Lr34/Yr18 haplotypes have been identified from sequencing the encoding ATP Binding Cassette transporter gene from diverse wheat germplasm of which one haplotype is commonly associated with the resistance phenotype. The paucity of well characterised APR genes, particularly for stem rust, calls for a focused effort in developing critical genetic stocks to delineate quantitative trait loci, construct specific BAC libraries for targeted APR genes to facilitate robust marker development for breeding applications, and the eventual cloning of the encoding genes.     

Development of simple PCR-based markers linked to the <Emphasis Type="Italic">Ms</Emphasis> locus,a restorer-of-fertility gene in onion (<Emphasis Type="Italic">Allium cepa</Emphasis> L.)

In order to implement reliable marker-assisted selection systems for the restorer-of-fertility locus (Ms) in onions (Allium cepa L.), simple PCR-based codominant markers linked to the Ms locus were developed. Based on the EST probe sequences of previously reported RFLP markers, full-length genomic sequences of the gene encoding putative oligopeptide transporter (OPT) was obtained by RACE. The first intron contained two 108 and 439-bp indel polymorphisms between the two Ms allele-linked OPT alleles. A simple PCR marker for OPT was developed by designing a primer pair on the flanking regions of the 108-bp indel which is created by two tandem repeats. The second simple PCR marker was developed from the EST probe encoding photosystem I subunit O (PsaO). Two 14 and 39-bp tandem repeats were identified from the 5′ upstream sequences of the PsaO-coding gene, which were isolated by genome walking. Three different compositions of these tandem repeats were identified from diverse onion germplasm. A primer set binding to the flanking sequence of these polymorphic repeats was used to amplify three different marker haplotypes. The OPT marker was tightly linked to the Ms locus at a distance of 1.5 cM, but the analysis of the linkage relationship showed little linkage disequilibrium between the marker and the Ms locus. Even so, these simple PCR markers are valuable tools for the marker-assisted selection of segregating individuals in onion F₁ hybrid breeding programs.