The accuracy of varietal selection using factor analytic models for multi-environment plant breeding trials

Modeling of cultivar x trial effects for multienvironment trials (METs) within a mixed model framework is now common practice in many plant breeding programs. The factor analytic (FA) model is a parsimonious form used to approximate the fully unstructured form of the genetic variance-covariance matrix in the model for MET data. In this study, we demonstrate that the FA model is generally the model of best fit across a range of data sets taken from early generation trials in a breeding program. In addition, we demonstrate the superiority of the FA model in achieving the most common aim of METs, namely the selection of superior genotypes. Selection is achieved using best linear unbiased predictions (BLUPs) of cultivar effects at each environment, considered either individually or as a weighted average across environments. In practice, empirical BLUPs (E-BLUPs) of cultivar effects must be used instead of BLUPs since variance parameters in the model must be estimated rather than assumed known. While the optimal properties of minimum mean squared error of prediction (MSEP) and maximum correlation between true and predicted effects possessed by BLUPs do not hold for E-BLUPs, a simulation study shows that E-BLUPs perform well in terms of MSEP.

Diel variation in mangrove fish abundances and trophic guilds of northeastern Australian estuaries with a proposed trophodynamic model

Diel activity patterns of tropical fish assemblages in turbid, mangrove-dominated estuaries remain largely undocumented, leading to uncertainty about ecological processes in these systems. To capture active fishes by day and night, gill nets were set perpendicular to mangrove shorelines, in six northeastern Australian estuaries during 13 bimonthly trips. Fish were sampled with eight large mesh (102-151 mm) nets, set for 6 hrs (1500-2100), and checked hourly (1146 day, 635 dusk, 872 night checks). Four smaller mesh (19-51 mm) nets were also set for 1 hr before and after sunset (77 day, 78 night checks). Of 157 total species, 22 were netted exclusively before sunset and 47 exclusively after sunset. All of the top 26 species were present both day and night, but of these, 46% were primarily nocturnal (diel index > 0.65). An average of 77.2 fish hr−1 were netted by day vs 171.4 by night. Within the 400 km coastal region, assemblages differed between two northern wave-dominated (WD) estuaries and four southern tide-dominated ('I'D) estuaries. In all six estuaries Lates calcarifer (Bloch, 1790) dominated night assemblages. In 'I'D estuaries, night assemblages were also dominated by Thryssa hamiltoni Gray, 1835 and Eleutheronema tetradactylum (Shaw, 1804); while in WD estuaries Herklotsichthys castelnaui (Ogilby, 1897), Leiognathus equulus (Forsskål, 1775), and Megalops cyprinoids (Broussonet, 1782) were dominant at night. Nocturnal species included planktivores and carnivores, while daytime assemblages were dominated by detritivores (Mugillidae). Higher night catch rates are attributed to increased activity by mobile fishes moving from mangrove to adjacent habitats to forage, especially immediately post-sunset. Although day-night diets and forage resources have yet to be compared in mangrove systems, previously unrecognized trophic relationships involving variation in diel activity among important fishery species (Centropomidae, polynemidae, Carangidae) and their prey may be key ecological processes in these tropical mangrove estuaries. A proposed hypothesis explaining diel variation in mangrove fish assemblages of tropical estuaries is presented through a conceptual model.

How accurate are the marker orders in crop linkage maps generated from large marker datasets?

Marker ordering during linkage map construction is a critical component of QTL mapping research. In recent years, high-throughput genotyping methods have become widely used, and these methods may generate hundreds of markers for a single mapping population. This poses problems for linkage analysis software because the number of possible marker orders increases exponentially as the number of markers increases. In this paper, we tested the accuracy of linkage analyses on simulated recombinant inbred line data using the commonly used Map Manager QTX (Manly et al. 2001: Mammalian Genome 12, 930-932) software and RECORD (Van Os et al. 2005: Theoretical and Applied Genetics 112, 30-40). Accuracy was measured by calculating two scores: % correct marker positions, and a novel, weighted rank-based score derived from the sum of absolute values of true minus observed marker ranks divided by the total number of markers. The accuracy of maps generated using Map Manager QTX was considerably lower than those generated using RECORD. Differences in linkage maps were often observed when marker ordering was performed several times using the identical dataset. In order to test the effect of reducing marker numbers on the stability of marker order, we pruned marker datasets focusing on regions consisting of tightly linked clusters of markers, which included redundant markers. Marker pruning improved the accuracy and stability of linkage maps because a single unambiguous marker order was produced that was consistent across replications of analysis. Marker pruning was also applied to a real barley mapping population and QTL analysis was performed using different map versions produced by the different programs. While some QTLs were identified with both map versions, there were large differences in QTL mapping results. Differences included maximum LOD and R-2 values at QTL peaks and map positions, thus highlighting the importance of marker order for QTL mapping

Losses of nitrogen in surface runoff from a plantation horticulture farm in coastal Queensland, Australia.

Surface losses of nitrogen from horticulture farms in coastal Queensland, Australia, may have the potential to eutrophy sensitive coastal marine habitats nearby. A case-study of the potential extent of such losses was investigated in a coastal macadamia plantation. Nitrogen losses were quantified in 5 consecutive runoff events during the 13-month study. Irrigation did not contribute to surface flows. Runoff was generated by storms at combined intensities and durations that were 20–40 mm/h for >9 min. These intensities and durations were within expected short-term (1 year) and long-term (up to 20 years) frequencies of rainfall in the study area. Surface flow volumes were 5.3 ± 1.1% of the episodic rainfall generated by such storms. Therefore, the largest part of each rainfall event was attributed to infiltration and drainage in this farm soil (Kandosol). The estimated annual loss of total nitrogen in runoff was 0.26 kg N/ha.year, representing a minimal loading of nitrogen in surface runoff when compared to other studies. The weighted average concentrations of total sediment nitrogen (TSN) and total dissolved nitrogen (TDN) generated in the farm runoff were 2.81 ± 0.77% N and 1.11 ± 0.27 mg N/L, respectively. These concentrations were considerably greater than ambient levels in an adjoining catchment waterway. Concentrations of TSN and TDN in the waterway were 0.11 ± 0.02% N and 0.50 ± 0.09 mg N/L, respectively. The steep concentration gradient of TSN and TDN between the farm runoff and the waterway demonstrated the occurrence of nutrient loading from the farming landscapes to the waterway. The TDN levels in the stream exceeded the current specified threshold of 0.2–0.3 mg N/L for eutrophication of such a waterway. Therefore, while the estimate of annual loading of N from runoff losses was comparatively low, it was evident that the stream catchment and associated agricultural land uses were already characterised by significant nitrogen loadings that pose eutrophication risks. The reported levels of nitrogen and the proximity of such waterways (8 km) to the coastline may have also have implications for the nearshore (oligotrophic) marine environment during periods of turbulent flow.

Separating grazing and rainfall effects at regional scale using remote sensing imagery: A dynamic reference-cover method

Remote detection of management-related trend in the presence of inter-annual climatic variability in the rangelands is difficult. Minimally disturbed reference areas provide a useful guide, but suitable benchmarks are usually difficult to identify. We describe a method that uses a unique conceptual framework to identify reference areas from multitemporal sequences of ground cover derived from Landsat TM and ETM+ imagery. The method does not require ground-based reference sites nor GIS layers about management. We calculate a minimum ground cover image across all years to identify locations of most persistent ground cover in years of lowest rainfall. We then use a moving window approach to calculate the difference between the window's central pixel and its surrounding reference pixels. This difference estimates ground-cover change between successive below-average rainfall years, which provides a seasonally interpreted measure of management effects. We examine the approach's sensitivity to window size and to cover-index percentiles used to define persistence. The method successfully detected management-related change in ground cover in Queensland tropical savanna woodlands in two case studies: (1) a grazing trial where heavy stocking resulted in substantial decline in ground cover in small paddocks, and (2) commercial paddocks where wet-season spelling (destocking) resulted in increased ground cover. At a larger scale, there was broad agreement between our analysis of ground-cover change and ground-based land condition change for commercial beef properties with different a priori ratings of initial condition, but there was also some disagreement where changing condition reflected pasture composition rather than ground cover. We conclude that the method is suitably robust to analyse grazing effects on ground cover across the 1.3 x 10(6) km(2) of Queensland's rangelands. Crown Copyright (c) 2012 Published by Elsevier Inc. All rights reserved.

Prediction of breeding values for average fruit weight in mango using a multivariate individual mixed model

Mango is an important horticultural fruit crop and breeding is a key strategy to improve ongoing sustainability. Knowledge of breeding values of potential parents is important for maximising progress from breeding. This study successfully employed a mixed linear model methods incorporating a pedigree to predict breeding values for average fruit weight from highly unbalanced data for genotypes planted over three field trials and assessed over several harvest seasons. Average fruit weight was found to be under strong additive genetic control. There was high correlation between hybrids propagated as seedlings and hybrids propagated as scions grafted onto rootstocks. Estimates of additive genetic correlation among trials ranged from 0.69 to 0.88 with correlations among harvest seasons within trials greater than 0.96. These results suggest that progress from selection for broad adaptation can be achieved, particularly as no repeatable environmental factor that could be used to predict G x E could be identified. Predicted breeding values for 35 known cultivars are presented for use in ongoing breeding programs.

Demonstration reaches: Looking back whilst moving forward with river rehabilitation under the Native Fish Strategy

‘Demonstration reaches’ are sections of river where multiple threats to native fish are addressed through river rehabilitation and strong community participation. They are an important way of promoting the key driving actions of the Murray-Darling Basin Authority's Native Fish Strategy (NFS) by using on-ground community-driven rehabilitation. Measuring rehabilitation success against well-defined targets and using this information to adaptively mange activities is fundamental to the demonstration reach philosophy. Seven years on from the establishment of the first demonstration reach, there are now seven throughout the Murray-Darling Basin (MDB), all in differing states of maturation and but all applying a standardised framework for monitoring native fish outcomes. In this study, we reflect on the role that demonstration reaches have played within the NFS, synthesise some key findings from 32 monitoring and evaluation outputs, and highlight some of the successes and barriers to success. We make recommendations as to how to strengthen the demonstration reach model to ensure it remains a relevant approach for fish habitat rehabilitation beyond the NFS and MDB.