ORCID Profile
0000-0002-3320-8938
Current Organisation
James Cook University
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
Publisher: Wiley
Date: 03-2008
DOI: 10.1111/J.1471-8286.2007.01940.X
Abstract: Previously reported maximum-likelihood pairwise relatedness (r) estimator of Thompson and Milligan (M) was extended to allow for negative r estimates under the regression interpretation of r. This was achieved by establishing the equivalency of the likelihoods used in the kinship program and the likelihoods of Thompson. The new maximum-likelihood (ML) estimator was evaluated by Monte Carlo simulations. It was found that the new ML estimator became unbiased significantly faster compared to the original M estimator when the amount of genotype information was increased. The effects of allele frequency estimation errors on the new and existing relatedness estimators were also considered.
Publisher: No publisher found
Date: 1993
Publisher: IEEE
Date: 12-2019
Publisher: IOP Publishing
Date: 28-11-1994
Publisher: IEEE
Date: 12-2018
Publisher: IOP Publishing
Date: 28-07-1994
Publisher: IOP Publishing
Date: 28-07-1994
Publisher: IEEE
Date: 12-2019
Publisher: Springer Science and Business Media LLC
Date: 14-02-2019
DOI: 10.1038/S41598-018-38343-3
Abstract: Robotic weed control has seen increased research of late with its potential for boosting productivity in agriculture. Majority of works focus on developing robotics for croplands, ignoring the weed management problems facing rangeland stock farmers. Perhaps the greatest obstacle to widespread uptake of robotic weed control is the robust classification of weed species in their natural environment. The unparalleled successes of deep learning make it an ideal candidate for recognising various weed species in the complex rangeland environment. This work contributes the first large, public, multiclass image dataset of weed species from the Australian rangelands allowing for the development of robust classification methods to make robotic weed control viable. The DeepWeeds dataset consists of 17,509 labelled images of eight nationally significant weed species native to eight locations across northern Australia. This paper presents a baseline for classification performance on the dataset using the benchmark deep learning models, Inception-v3 and ResNet-50. These models achieved an average classification accuracy of 95.1% and 95.7%, respectively. We also demonstrate real time performance of the ResNet-50 architecture, with an average inference time of 53.4 ms per image. These strong results bode well for future field implementation of robotic weed control methods in the Australian rangelands.
Publisher: American Physical Society (APS)
Date: 06-1991
Publisher: American Physical Society (APS)
Date: 09-11-2012
Publisher: No publisher found
Date: 1994
Publisher: Wiley
Date: 10-09-2020
DOI: 10.1111/FAF.12501
Publisher: ACM
Date: 16-06-2018
Publisher: IOP Publishing
Date: 28-05-1993
Publisher: American Physical Society (APS)
Date: 12-1991
Publisher: American Physical Society (APS)
Date: 13-08-2010
Publisher: Springer Science and Business Media LLC
Date: 12-10-2018
Publisher: Elsevier BV
Date: 08-2015
Publisher: IOP Publishing
Date: 28-04-1991
Publisher: American Chemical Society (ACS)
Date: 30-06-2007
DOI: 10.1021/CI700100F
Abstract: Using the largest available database of 328 blood-brain distribution (logBB) values, a quantitative benchmark was proposed to allow for a consistent comparison of the predictive accuracy of current and future logBB/quantitative structure-activity relationship (-QSAR) models. The usefulness of the benchmark was illustrated by comparing the global and k-nearest neighbors (kNN) multiple-linear regression (MLR) models based on the linear free-energy relationship (LFER) descriptors, and one non-LFER-based MLR model. The leave-one-out (LOO) and leave-group-out Monte Carlo (MC) cross-validation results (q(2) = 0.766, qms = 0.290, and qms(mc) = 0.311) indicated that the LFER-based kNN-MLR model was currently one of the most accurate predictive logBB-QSAR models. The LOO, MC, and kNN-MLR methods have been implemented in the QSAR-BENCH program, which is freely available from www.dmitrykonovalov.org for academic use.
Publisher: American Physical Society (APS)
Date: 02-1991
Publisher: Elsevier BV
Date: 03-2022
DOI: 10.1016/J.MARENVRES.2022.105568
Abstract: High quality nursery grounds are important for species success and the long-term sustainability of fish stocks. However, even for important fisheries species, what constitutes nursery habitats is only coarsely defined, and details of specific requirements are often lacking. In this study we investigated upstream estuarine areas in central Queensland, Australia, to identify the environmental factors that constrain nursery ground utilization for important fisheries species. We used unbaited underwater video cameras to assess fish presence, and used a range of water quality sensors to record fluctuations in environmental conditions, likely to influence juveniles, over several months (e.g. tidal connection patterns, temperature, salinity and dissolved oxygen). We found that juveniles of three fisheries target species (Lutjanus argentimaculatus, Lutjanus russellii and Acanthopagrus australis) were common in the upstream sections of the estuaries. For each species, only a subset of the factors assessed were influential in determining nursery ground utilization, and their importance varied among species, even among the closely related L. argentimaculatus and L. russellii. Overall, tidal connectivity and the availability of complex structure, were the most influential factors. The reasons for the importance of connectivity are complex as well as allowing access, tidal connectivity influences water levels, water temperature and dissolved oxygen - all important physiological requirements for successful occupation. The impact of variation in juvenile access to food and refuge in nursery habitat was not directly assessed. While crucial, these factors are likely to be subordinate to the suite of environmental characteristics necessary for the presence and persistence of juveniles in these locations. These results suggest that detailed environmental and biological knowledge is necessary to define the nuanced constraints of nursery ground value among species, and this detailed knowledge is vital for informed management of early life-history stages.
Publisher: IEEE
Date: 07-2019
Publisher: IOP Publishing
Date: 12-01-2012
Publisher: American Physical Society (APS)
Date: 12-09-2011
Publisher: Elsevier
Date: 2010
Publisher: MDPI AG
Date: 09-04-2020
DOI: 10.3390/INFO11040200
Abstract: A predictable aggregation of dwarf minke whales (Balaenoptera acutorostrata subspecies) occurs annually in the Australian waters of the northern Great Barrier Reef in June–July, which has been the subject of a long-term photo-identification study. Researchers from the Minke Whale Project (MWP) at James Cook University collect large volumes of underwater digital imagery each season (e.g., 1.8TB in 2018), much of which is contributed by citizen scientists. Manual processing and analysis of this quantity of data had become infeasible, and Convolutional Neural Networks (CNNs) offered a potential solution. Our study sought to design and train a CNN that could detect whales from video footage in complex near-surface underwater surroundings and differentiate the whales from people, boats and recreational gear. We modified known classification CNNs to localise whales in video frames and digital still images. The required high classification accuracy was achieved by discovering an effective negative-labelling training technique. This resulted in a less than 1% false-positive classification rate and below 0.1% false-negative rate. The final operation-version CNN-pipeline processed all videos (with the interval of 10 frames) in approximately four days (running on two GPUs) delivering 1.95 million sorted images.
Publisher: IEEE
Date: 12-2019
Publisher: IOP Publishing
Date: 14-09-1992
Publisher: American Physical Society (APS)
Date: 11-1991
Publisher: Springer Science and Business Media LLC
Date: 10-2017
Publisher: AIP Publishing
Date: 25-03-2015
DOI: 10.1063/1.4915888
Abstract: We report differential cross sections (DCSs) for electron-impact vibrational-excitation of tetrahydrofuran, at intermediate incident electron energies (15-50 eV) and over the 10°-90° scattered electron angular range. These measurements extend the available DCS data for vibrational excitation for this species, which have previously been obtained at lower incident electron energies (≤20 eV). Where possible, our data are compared to the earlier measurements in the overlapping energy ranges. Here, quite good agreement was generally observed where the measurements overlapped.
Publisher: Springer Science and Business Media LLC
Date: 30-05-2013
DOI: 10.1007/S10126-013-9514-3
Abstract: Pearl oysters are not only farmed for their gemstone quality pearls worldwide, but they are also becoming important model organisms for investigating genetic mechanisms of biomineralisation. Despite their economic and scientific significance, limited genomic resources are available for this important group of bivalves, h ering investigations into identifying genes that regulate important pearl quality traits and unique biological characteristics (i.e. biomineralisation). The silver-lipped pearl oyster, Pinctada maxima, is one species where there is interest in understanding genes that regulate commercially important pearl traits, but presently, there is a dearth of genomic information. The objective of this study was to develop and validate a large number of type I genome-wide single nucleotide polymorphisms (SNPs) for P. maxima suitable for high-throughput genotyping. In addition, sequence annotations and Gene Ontology terms were assigned to a large mantle tissue 454 expressed sequence tag assembly (96,794 contigs) and information on known bivalve biomineralisation genes was incorporated into SNP discovery. The SNP discovery effort resulted in the de novo identification of 172,625 SNPs, of which 9,108 were identified as high value [minor allele frequency (MAF)≥ 0.15, read depth ≥ 8]. Validation of 2,782 of these SNPs using Illumina iSelect Infinium genotyping technology returned some of the highest assay conversion (86.6 %) and validation (59.9 % mean MAF 0.28) rates observed in aquaculture species to date. Genomic resources presented here will be pivotal to future research investigating the biological mechanisms behind biomineralisation and will form a strong foundation for genetic selective breeding programs in the P. maxima pearling industry.
Publisher: AIP Publishing
Date: 25-03-2015
DOI: 10.1063/1.4915889
Abstract: In this paper, we report newly derived integral cross sections (ICSs) for electron impact vibrational excitation of tetrahydrofuran (THF) at intermediate impact energies. These cross sections extend the currently available data from 20 to 50 eV. Further, they indicate that the previously recommended THF ICS set [Garland et al., Phys. Rev. A 88, 062712 (2013)] underestimated the strength of the electron-impact vibrational excitation processes. Thus, that recommended vibrational cross section set is revised to address those deficiencies. Electron swarm transport properties were calculated with the amended vibrational cross section set, to quantify the role of electron-driven vibrational excitation in describing the macroscopic swarm phenomena. Here, significant differences of up to 17% in the transport coefficients were observed between the calculations performed using the original and revised cross section sets for vibrational excitation.
Publisher: Scientific Research Publishing, Inc.
Date: 2018
Publisher: ACM
Date: 25-08-2017
Publisher: Wiley
Date: 11-10-2004
Publisher: American Physical Society (APS)
Date: 10-1994
Publisher: Oxford University Press (OUP)
Date: 03-03-2005
DOI: 10.1093/BIOINFORMATICS/BTI373
Abstract: Accuracy testing of various pedigree reconstruction methods requires an efficient algorithm for the calculation of distance between a known partition and its reconstruction. The currently used algorithm of Almudevar and Field takes a prohibitively long time for certain partitions and population sizes. We present an algorithm that very efficiently reduces the partition-distance calculation to the classic assignment problem of weighted bipartite graphs that has known polynomial-time solutions. The performance of the algorithm is tested against the Almudevar and Field partition-distance algorithm to verify the significant improvement in speed. Computer code written in java is available upon request from the first author.
Publisher: American Chemical Society (ACS)
Date: 10-2008
DOI: 10.1021/CI800209K
Abstract: A quantitative structure-activity relationship (QSAR) model is typically developed to predict the biochemical activity of untested compounds from the compounds' molecular structures. "The gold standard" of model validation is the blindfold prediction when the model's predictive power is assessed from how well the model predicts the activity values of compounds that were not considered in any way during the model development/calibration. However, during the development of a QSAR model, it is necessary to obtain some indication of the model's predictive power. This is often done by some form of cross-validation (CV). In this study, the concepts of the predictive power and fitting ability of a multiple linear regression (MLR) QSAR model were examined in the CV context allowing for the presence of outliers. Commonly used predictive power and fitting ability statistics were assessed via Monte Carlo cross-validation when applied to percent human intestinal absorption, blood-brain partition coefficient, and toxicity values of saxitoxin QSAR data sets, as well as three known benchmark data sets with known outlier contamination. It was found that (1) a robust version of MLR should always be preferred over the ordinary-least-squares MLR, regardless of the degree of outlier contamination and that (2) the model's predictive power should only be assessed via robust statistics. The Matlab and java source code used in this study is freely available from the QSAR-BENCH section of www.dmitrykonovalov.org for academic use. The Web site also contains the java-based QSAR-BENCH program, which could be run online via java's Web Start technology (supporting Windows, Mac OSX, Linux/Unix) to reproduce most of the reported results or apply the reported procedures to other data sets.
Publisher: Oxford University Press (OUP)
Date: 23-08-2005
DOI: 10.1093/BIOINFORMATICS/BTI642
Abstract: The problem of reconstructing full sibling groups from DNA marker data remains a significant challenge for computational biology. A recently published heuristic algorithm based on Mendelian exclusion rules and the Simpson index was successfully applied to the full sibship reconstruction (FSR) problem. However, the so-called SIMPSON algorithm has an unknown complexity measure, questioning its applicability range. We present a modified version of the SIMPSON (MS) algorithm that behaves as O(n(3)) and achieves the same or better accuracy when compared with the original algorithm. Performance of the MS algorithm was tested on a variety of simulated diploid population s les to verify its complexity measure and the significant improvement in efficiency (e.g. 100 times faster than SIMPSON in some cases). It has been shown that, in theory, the SIMPSON algorithm runs in non-polynomial time, significantly limiting its usefulness. It has been also verified via simulation experiments that SIMPSON could run in O(n(a)), where a > 3. Computer code written in Java is available upon request from the first author. Dmitry.Konovalov@jcu.edu.au.
Publisher: Wiley
Date: 19-10-2021
DOI: 10.1111/EVO.14348
Publisher: American Chemical Society (ACS)
Date: 31-01-2008
DOI: 10.1021/CI700283S
Abstract: A new variable selection wrapper method named the Monte Carlo variable selection (MCVS) method was developed utilizing the framework of the Monte Carlo cross-validation (MCCV) approach. The MCVS method reports the variable selection results in the most conventional and common measure of statistical hypothesis testing, the P-values, thus allowing for a clear and simple statistical interpretation of the results. The MCVS method is equally applicable to the multiple-linear-regression (MLR)-based or non-MLR-based quantitative structure-activity relationship (QSAR) models. The method was applied to blood-brain barrier (BBB) permeation and human intestinal absorption (HIA) QSAR problems using MLR to demonstrate the workings of the new approach. Starting from more than 1600 molecular descriptors, only two (TPSA(NO) and ALOGP) yielded acceptably low P-values for the BBB and HIA problems, respectively. The new method has been implemented in the QSAR-BENCH v2 program, which is freely available (including its Java source code) from www.dmitrykonovalov.org for academic use.
Publisher: Scientific Research Publishing, Inc.
Date: 2018
Publisher: IEEE
Date: 12-2019
Publisher: American Physical Society (APS)
Date: 12-1991
Publisher: IOP Publishing
Date: 14-03-1995
Publisher: EDP Sciences
Date: 06-09-2008
DOI: 10.1051/ITA:2007034
Publisher: IOP Publishing
Date: 15-05-2018
No related grants have been discovered for Dmitry Konovalov.