Validation Sample Set Research Articles

Comparisons of three recent moderate resolution African land cover datasets: CGLS-LC100, ESA-S2-LC20, and FROM-GLC-Africa30

ABSTRACTAccurate and up-to-date land use land cover (LULC) mapping has long been a challenge in Africa. Recently, three LULC maps with moderate spatial resolution (20 m to 100 m) have been developed using multiple Earth observation datasets for 2015–2016 for the whole continent, which provide unprecedented spatial detail of the land surface for Africa. This study aimed to compare these three recent African LULC maps (i.e. the Copernicus Global Land Service Land Cover (CGLS-LC100, 100 m), European Space Agency Sentinel-2A Land Cover (ESA-S2-LC20, 20 m) and Finer Resolution Observation and Monitoring of Global Land Cover for Africa version 2 (FROM-GLC-Africa30, 30 m)) using a validation sample set and statistics from the FAO. The results indicated that the accuracy of the three datasets was unevenly distributed in spatial extent and area estimation. All the three datasets achieve an accuracy of above 60% and the fraction layer of CGLS-LC100 showed the best consistency with FAO statistics in the area. However, great disagreement in spatial details was found among three products, with 43.12% of the total area in Africa was in low agreement. The LULC mapping regions with the highest uncertainty were southeast Africa, the Sahel region and the Eastern Africa Plateau. Uncertainty was most closely related to elevation and precipitation changes along latitude/longitude.

Identification of Multisialylated LacdiNAc Structures as Highly Prostate Cancer Specific Glycan Signatures on PSA.

Serum prostate-specific antigen (PSA) test is the current gold standard for screening and diagnosis of prostate cancer (PCa), while overdiagnosis and overtreatment are social problems. In order to improve the specificity and exclude a false positive diagnosis in PSA test, PCa-specific glycosylation subtypes of PSA were explored using in-depth quantitative profiling of PSA glycoforms based on mass spectrometric oxonium ion monitoring technology. As a result of analysis using sera from 15 PCa or 15 benign prostate hyperplasia (BPH) patients whose PSA levels were in the "gray zone" (4.0-10.0 ng/mL), 52 glycan structures on PSA were quantitatively observed. We found that abundance of multisialylated LacdiNAc (GalNAcβ1-4GlcNAc) structures were significantly upregulated in the PCa group compared to the BPH group. A couple of those glycoforms were then extracted and subjected to establish a novel PCa-specific diagnosis model (PSA G-index). When the diagnostic power was assessed using an independent validation sample set (15 PCa and 15 BPH patients in the PSA gray zone), an AUC of PSA G-index was 1.00, while that of total PSA or PSA f/T ratio was 0.50 or 0.60, respectively. Moreover, both PSA glycoforms showed significant correlation with Gleason scores. Lectin histochemical staining analysis also showed that PCa cells overexpressed glycoproteins containing LacdiNAc and sialic acids moieties. Thus, PSA G-index could serve as not only an effective secondary screening method to exclude false positive diagnosis in PSA screening, but also a potential grading biomarker for PCa.

Comparative non-destructive classification of partial waxy wheats using near-infrared and Raman spectroscopy

Waxy proteins play a key role in amylose synthesis in wheat. Eight lines of common wheat (Triticum aestivum L.) carrying mutations in the three homoeologous waxy loci, Wx-A1, Wx-B1 and Wx-D1, have been classified by near-infrared (NIR) and Raman spectroscopy combined with chemometrics. Sample spectra from wheat seeds were collected by using a NIR spectrometer in the wave rage 1600–2400 nm, and then Raman spectrometer in the wave range 700–2000 cm–1. All samples were split randomly into a calibration sample set containing 284 seeds (~35 seeds per line) and a validation sample set containing the remaining 92 seeds. Classification of these samples was undertaken by discriminant analysis combined with principal component analysis (PCA) based on the raw spectra processed by appropriate pre-treatment methods. The classification results by discriminant analysis indicated that the percentage of correctly identified samples by NIR spectroscopy was 84.2% for the calibration set and 84.8% for the validation set, and by Raman spectroscopy 94.4% and 94.6%, respectively. The results demonstrated that Raman spectroscopy combined with chemometrics as a rapid method is superior to NIR spectroscopy in classifying eight partial waxy wheat lines with different waxy proteins.

Rapid Detection of Live and Dead Escherichia coli in a Suspension Using Spectroscopy and Chemometrics

Viability assessment of bacteria is critical in monitoring of food or environmental samples. Existing methods are time-consuming, labour-intensive or require trained manpower and costly chemicals. Potential of commonly used UV-visual spectrometer was explored for rapid viability detection of Escherichia coli (ATCC 8739). Spectra of samples (live and dead cells) mixed in different proportion revealed clear differences. Live bacterial suspension showed absorption peak at 260 nm with decreasing amplitude as the proportion of live bacteria was reduced in the suspension and vice-versa. Principal component analyses (PCA) of spectral data showed clear clustering of samples based on the level of live bacterial cells (5 % significance level). Soft Independent Modelling of Class Analogy (SIMCA) and Partial Least Square Discriminant Analysis (PLSDA) yielded 100 % correct classification with test samples. The percentage of live and dead bacteria in a suspension could be predicted with coefficient of determination (R2 ) of 0.980 and 0.977 for calibration and validation sample sets, respectively, in the range of 259-261 nm using Multiple Linear Regression (MLR). Low standard errors of calibration (4.5), prediction (4.8) and high R2 (0.98) indicated the potential of UV visual spectrometer to detect and predict live and dead cells of E. coli in a suspension.

Nonlinear Modeling of the Flux Linkage in 2-D Plane for the Planar Switched Reluctance Motor

This paper proposes a nonlinear flux linkage model in 2-D plane for the planar switched reluctance motor (PSRM). The inputs of the proposed model are the 2-D positions and the current, and the output is the flux linkage. The proposed model is established via a cascade-forward backpropagation neural network (CFNN). The designed CFNN consists of four layers: one input layer, two hidden layers, and one output layer. The first hidden layer has 20 neurons with a tan-sigmoid transfer function, and the second hidden layer has 20 neurons with a log-sigmoid transfer function. The output layer is a pure linear layer. The sample set with 179 755 samples is obtained experimentally in a dSPACE-based PSRM system by applying the dc excitation method. The sample set is divided into three sets. 35% and 30% of the samples are randomly chosen as the training sample set and validation sample set, respectively, and the remaining samples are utilized as the test sample set to assess the generalization performance of the CFNN-based model. According to the results of the test sample set, the maximum relative error is 11.05% and the mean relative error is 0.42% when the current ranges from 1 to 9 A. The CFNN has the capability to build a multi-input nonlinear model. The CFNN-based model is capable of reflecting the variations of flux linkage in 2-D plane caused by manufacturing tolerances. The effectiveness of the CFNN-based model is finally verified.

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) proteomic profiling of cerebrospinal fluid in the diagnosis of enteroviral meningitis: a proof-of-principle study.

The use of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) for diagnosing viral infections by directly testing clinical specimens has not previously been explored. In this proof-of-principle study, we tested the hypothesis that proteomic profiling of cerebrospinal fluid (CSF) by mass spectrometry may be useful in the diagnosis of enteroviral (EV) meningitis. A total of 114 cryopreserved CSF samples were analyzed, of which 47 were positive for EV and 67 were negative. Total CSF proteins were precipitated and subjected to MALDI-TOF-MS analysis in a low (2-20kDa) molecular weight range using a MicroFlex LT mass spectrometer. The whole data set was randomly split into a training set (n = 76 specimens) and a validation set (n = 38 samples). Backward/forward stepwise logistic regression analyses identified 30 peaks that were differentially present in EV-positive and EV-negative specimens. These were used to build a model which displayed an overall classification accuracy of 93%. The discriminative ability of the model was confirmed by using a validation sample set (overall accuracy 83%). In fact, the model was able to correctly classify 61 out of 67 EV-negative samples and 42 out of 47 EV-positive specimens. EV meningitis is associated with a distinctive protein profile that may be directly detectable in CSF specimens by MALDI-TOF-MS.

Mapping annual urban dynamics (1985–2015) using time series of Landsat data

The information of urban dynamics at fine spatiotemporal resolutions is crucial to urban growth modeling and sustainable urban development. However, there are still challenges in deriving the change information of urbanization in timing and location over a long period. In this study, we developed a framework to map urban expansion at an annual interval from 1985 to 2015 by using the time series of Landsat data. First, the time series of Landsat data (1985–2015) were grouped into three periods, i.e., 1985–2001, 2001–2011, and 2011–2015, according to the available National Land Cover Database (NLCD). Then, a temporal segmentation approach was implemented for each period using three indicators representing changes from vegetation, water, and bare land to urban. Turning years of the start and end of change were identified. Three temporal segments representing phases of prior change, change, and post change, were generated accordingly. Thereafter, urban extents before 2001 and after 2011 were classified using a change vector analysis (CVA) based approach aided by the NLCD and identified temporal segments. Finally, urbanized pixels in each period were determined according to the identified turning years. Our approach of temporal segmentation is reliable for detecting changes caused by urban growth, with an overall accuracy of 90% in identifying turning years (±1 year). Using an independent validation sample set, the CVA based approach reaches an overall accuracy of 87%. The derived product of urban dynamics shows a relatively stable increment of urban growth in Des Moines and Ames, Iowa, US, and most urbanized areas were converted from vegetated lands within 2–3 years. The proposed framework is capable of mapping long-term dynamics of urban extents at an annual interval and the outcome is useful in effectively updating current products of urban extents and improving urban growth modeling.

Distinctiveness, complexity, and repeatability of online signature templates

This paper proposes three measures to quantify the characteristics of online signature templates in terms of distinctiveness, complexity and repeatability. A distinctiveness measure of a signature template is computed from a set of enrolled signature samples and a statistical assumption about random signatures. Secondly, a complexity measure of the template is derived from a set of enrolled signature samples. Finally, given a signature template, a measure to quantify the repeatability of the online signature is derived from a validation set of samples. These three measures can then be used as an indicator for the performance of the system in rejecting random forgery samples and skilled forgery samples and the performance of users in providing accepted genuine samples, respectively. The effectiveness of these three measures and their applications are demonstrated through experiments performed on three online signature datasets and one keystroke dynamics dataset using different verification algorithms.

Rapid qualitative and quantitative analysis of methamphetamine, ketamine, heroin, and cocaine by near-infrared spectroscopy

Rapid and nondestructive near infrared spectroscopy (NIR) methods have been developed for simultaneous qualitative and quantitative analysis of methamphetamine, ketamine, heroin, and cocaine in seized samples. This is the first systematic report regarding a qualitative and quantitative procedure of applying NIR for drug analysis. A total of 282 calibration samples and 836 prediction samples were used for the building and validating of qualitative and quantitative models. Two qualitative analysis modeling methods for soft independent modeling by class analogy (SIMCA) and supporting vector machine (SVM) were compared. From its excellent performance in rejecting false positive results, SIMCA was chosen. The drug concentrations in the calibration and validation sample sets were analyzed using high-performance liquid chromatography. Based on the use of first-order derivative spectral data after standard normal variate (SNV) transformation correction, in the wavelength range from 10,000 to 4000cm−1, four partial least squares quantitative-analysis models were built. The coefficients of determination for all calibration models were >99.3, and the RMSEC, RMSECV, and RMSEP were all less than 1.6, 2.9, and 3.6%, respectively. The results obtained here indicated that NIR with chemometric methods was accurate for qualitative and quantitative analysis of drug samples. This methodology provided a potentially useful alternative to time-consuming gas chromatography–mass spectroscopy and high-performance liquid chromatography methods.

Determination of salvianolic acid B and borneol in compound Danshen tablet by near-infrared spectroscopy and establishment of dependency model

To determine the contents of salvianolic acid B and borneol in compound Danshen tablet (composed of Salviae Miltiorrhizae Radix et Rhizome, Notoginseng Radix et Rhizome and Borneolum Syntheticum) by near-infrared spectroscopy (NIR) and to establish a dependency model for rapid quantitative analysis. NIR data of 74 batches of compound Danshen tablet from different companies were collected; the contents of salvianolic acid B and borneol were determined by using high performanceliquid chromatography (HPLC) and gas chromatography (GC) respectively to establish the dependency model for salvianolic acid B and borneol. The results showed that the best waveband for salvianolic acid B was 10 846.2-10 013, 9 195.3-8 362.2, 6 719.1-4 242.8 cm⁻¹; root-mean-squares error of cross-validation (RMSECV), coefficient of determination R² and regression point displacement (RPD) of salvianolic acid B were 1.72 mg·g⁻¹, 91.05% and 7.93 mg·g⁻¹, respectively. While the best wavebands, RMSECV, R² and RPD of borneol were 10 846.2-5 060.5 cm⁻¹, 1.2 mg·g⁻¹, 96.11% and 6.71 mg·g⁻¹，respectively. The relative error of the established model as validation results for validation set samples was 2.67% and 4.64%， respectively.With the good predictability, the models can be applied to the real time monitoring and quality control of compound Danshen tablet.

Sequencing-based counting and size profiling of plasma Epstein–Barr virus DNA enhance population screening of nasopharyngeal carcinoma

Circulating tumor-derived DNA testing for cancer screening has recently been demonstrated in a prospective study on identification of nasopharyngeal carcinoma (NPC) among 20,174 asymptomatic individuals. Plasma EBV DNA, a marker for NPC, was detected using real-time PCR. While plasma EBV DNA was persistently detectable in 97.1% of the NPCs identified, ∼5% of the general population had transiently detectable plasma EBV DNA. We hypothesized that EBV DNA in plasma of subjects with or without NPC may have different molecular characteristics. We performed target-capture sequencing of plasma EBV DNA and identified differences in the abundance and size profiles of EBV DNA molecules within plasma of NPC and non-NPC subjects. NPC patients had significantly higher amounts of plasma EBV DNA, which showed longer fragment lengths. Cutoff values were established from an exploratory dataset and tested in a validation sample set. Adopting an algorithm that required a sample to concurrently pass cutoffs for EBV DNA counting and size measurements, NPCs were detected at a positive predictive value (PPV) of 19.6%. This represented superior performance compared with the PPV of 11.0% in the prospective screening study, which required participants with an initially detectable plasma EBV DNA result to be retested within 4 weeks. The observed differences in the molecular nature of EBV DNA molecules in plasma of subjects with or without NPC were successfully translated into a sequencing-based test that had a high PPV for NPC screening and achievable through single time-point testing.

Identification of Adulterated and Non-adulterated Norwegian Salmon Using FTIR and an Improved PLS-DA Method

Norwegian salmon is often adulterated with Heilongjiang salmon at local fish markets. To promote fair price competition at fish markets and protect consumer rights, we developed a quick and accurate identification method to distinguish adulterated and non-adulterated Norwegian salmon using a Fourier transform infrared spectroscopy (FTIR). In this study, Norwegian and Heilongjiang salmon could be readily distinguished using partial least squares discriminant analysis (PLS-DA), but it failed to detect the accurate level of 20 to 80% of adulterated Norwegian salmon samples. In order to improve the PLS-DA model, several pre-processing methods, including standard normal variate (SNV), multiplicative scatter correction (MSC), and normalization, were used to evaluate individually to select the most appropriate correction method. Characteristics of the spectra within the waveband range covering 450 to 4000 cm−1 were also analyzed to determine the optimum sub-waveband range to improve the accuracy of the model. The results of the study showed that using FTIR and the improved PLS-DA model established in this study, the adulterated and non-adulterated Norwegian salmon could be completely distinguished. The accuracy of the adulteration level and the prediction accuracy of the model were also significantly improved when normalization method was used at 450 to 1790 cm−1 sub-wavebands. For the calibration and cross-validation sample sets, the determination coefficients of the improved PLS-DA model were at 0.99 and 0.98, respectively. The mean square errors were 2.3 and 4%, resulting in a 90% accuracy of validation sample sets. This technology should provide fish markets an easy and reliable way to distinguish the adulterated and non-adulterated salmon.

Schizophrenia-risk and urban birth are associated with proteomic changes in neonatal dried blood spots

In the present study, we tested whether there were proteomic differences in blood between schizophrenia patients after the initial onset of the disorder and controls; and whether those differences were also present at birth among neonates who later developed schizophrenia compared to those without a psychiatric admission. We used multiple reaction monitoring mass spectrometry to quantify 77 proteins (147 peptides) in serum samples from 60 first-onset drug-naive schizophrenia patients and 77 controls, and 96 proteins (152 peptides) in 892 newborn blood-spot (NBS) samples collected between 1975 and 1985. Both serum and NBS studies showed significant alterations in protein levels. Serum results revealed that Haptoglobin and Plasma protease C1 inhibitor were significantly upregulated in first-onset schizophrenia patients (corrected P < 0.05). Alpha-2-antiplasmin, Complement C4-A and Antithrombin-III were increased in first-onset schizophrenia patients (uncorrected P-values 0.041, 0.036 and 0.013, respectively) and also increased in newborn babies who later develop schizophrenia (P-values 0.0058, 0.013 and 0.044, respectively). We also tested whether protein abundance at birth was associated with exposure to an urban environment during pregnancy and found highly significant proteomic differences at birth between urban and rural environments. The prediction model for urbanicity had excellent predictive performance in both discovery (area under the receiver operating characteristic curve (AUC) = 0.90) and validation (AUC = 0.89) sample sets. We hope that future biomarker studies based on stored NBS samples will identify prognostic disease indicators and targets for preventive measures for neurodevelopmental conditions, particularly those with onset during early childhood, such as autism spectrum disorder.

Calibration transfer of a Raman spectroscopic quantification method for the assessment of liquid detergent compositions from at-line laboratory to in-line industrial scale

Calibration transfer or standardisation aims at creating a uniform spectral response on different spectroscopic instruments or under varying conditions, without requiring a full recalibration for each situation. In the current study, this strategy is applied to construct at-line multivariate calibration models and consequently employ them in-line in a continuous industrial production line, using the same spectrometer.Firstly, quantitative multivariate models are constructed at-line at laboratory scale for predicting the concentration of two main ingredients in hard surface cleaners. By regressing the Raman spectra of a set of small-scale calibration samples against their reference concentration values, partial least squares (PLS) models are developed to quantify the surfactant levels in the liquid detergent compositions under investigation. After evaluating the models performance with a set of independent validation samples, a univariate slope/bias correction is applied in view of transporting these at-line calibration models to an in-line manufacturing set-up. This standardisation technique allows a fast and easy transfer of the PLS regression models, by simply correcting the model predictions on the in-line set-up, without adjusting anything to the original multivariate calibration models.An extensive statistical analysis is performed in order to assess the predictive quality of the transferred regression models. Before and after transfer, the R2 and RMSEP of both models is compared for evaluating if their magnitude is similar. T-tests are then performed to investigate whether the slope and intercept of the transferred regression line are not statistically different from 1 and 0, respectively. Furthermore, it is inspected whether no significant bias can be noted. F-tests are executed as well, for assessing the linearity of the transfer regression line and for investigating the statistical coincidence of the transfer and validation regression line. Finally, a paired t-test is performed to compare the original at-line model to the slope/bias corrected in-line model, using interval hypotheses.It is shown that the calibration models of Surfactant 1 and Surfactant 2 yield satisfactory in-line predictions after slope/bias correction. While Surfactant 1 passes seven out of eight statistical tests, the recommended validation parameters are 100% successful for Surfactant 2. It is hence concluded that the proposed strategy for transferring at-line calibration models to an in-line industrial environment via a univariate slope/bias correction of the predicted values offers a successful standardisation approach.

Irregular analytical errors in diagnostic testing - a novel concept.

In laboratory medicine, routine periodic analyses for internal and external quality control measurements interpreted by statistical methods are mandatory for batch clearance. Data analysis of these process-oriented measurements allows for insight into random analytical variation and systematic calibration bias over time. However, in such a setting, any individual sample is not under individual quality control. The quality control measurements act only at the batch level. Quantitative or qualitative data derived for many effects and interferences associated with an individual diagnostic sample can compromise any analyte. It is obvious that a process for a quality-control-sample-based approach of quality assurance is not sensitive to such errors. To address the potential causes and nature of such analytical interference in individual samples more systematically, we suggest the introduction of a new term called the irregular (individual) analytical error. Practically, this term can be applied in any analytical assay that is traceable to a reference measurement system. For an individual sample an irregular analytical error is defined as an inaccuracy (which is the deviation from a reference measurement procedure result) of a test result that is so high it cannot be explained by measurement uncertainty of the utilized routine assay operating within the accepted limitations of the associated process quality control measurements. The deviation can be defined as the linear combination of the process measurement uncertainty and the method bias for the reference measurement system. Such errors should be coined irregular analytical errors of the individual sample. The measurement result is compromised either by an irregular effect associated with the individual composition (matrix) of the sample or an individual single sample associated processing error in the analytical process. Currently, the availability of reference measurement procedures is still highly limited, but LC-isotope-dilution mass spectrometry methods are increasingly used for pre-market validation of routine diagnostic assays (these tests also involve substantial sets of clinical validation samples). Based on this definition/terminology, we list recognized causes of irregular analytical error as a risk catalog for clinical chemistry in this article. These issues include reproducible individual analytical errors (e.g. caused by anti-reagent antibodies) and non-reproducible, sporadic errors (e.g. errors due to incorrect pipetting volume due to air bubbles in a sample), which can both lead to inaccurate results and risks for patients.

Strategy for sensitive and specific somatic SNV/InDel calling: Leveraging different callers with in-house generated gold standard variant sets.

e13104 Background: Somatic mutation calling is critical for cancer genotyping. Although rapid development of mutation detection is witnessed with the maturity of NGS, the need for high sensitivity often results in compromised specificity and manual inspection. Here, we propose a methodology that leverages different variant callers to account for specificity without compromising sensitivity. Methods: We designed a cohort of training samples (n = 22), each with known set of SNVs/InDels that were discovered by KEW CANCERPLEX platform. We assessed the performance of four prevailing mutation callers that utilize different statistical approaches and therefore have different calls, using the training samples. We optimized the parameters of the four variant callers to detect all expected variants. We then examined the intersections of every combination of the four callers and identified the best one that eliminated the highest rate of false calls. A customized tool was developed for the intersection of component SNVs and InDels and the report differences among different callers. We also collected another set of validation samples (n = 28), each with true mutations that were both curated and orthogonally validated. We used this set to further test the efficacy of the refined calling strategy. Results: From the training samples, we chose the combination that provides highest sensitivity and specificity. This refined mutation calling strategy removed ~20% false SNV calls and ~50% false InDel calls in average for each sample. This result was further confirmed by the validation samples. Conclusions: The curated set of mutations from the genetic test platforms can provide valuable gold standard to test and tune mutation callers. The conclusions drawn from the curated variants are in line with the experimentally validated variants and showcase the validity of this practice, which we applied to demonstrate that the strategy of intersecting variants from optimized variant callers will generate mutation calls of higher specificity without compromising sensitivity. This methodology reduces the number of variants for curation and improves curation procedures and turn-around time.

Multiway analysis methods applied to the fluorescence excitation-emission dataset for the simultaneous quantification of valsartan and amlodipine in tablets

In this study, excitation-emission matrix datasets, which have strong overlapping bands, were processed by using four different chemometric calibration algorithms consisting of parallel factor analysis, Tucker3, three-way partial least squares and unfolded partial least squares for the simultaneous quantitative estimation of valsartan and amlodipine besylate in tablets. In analyses, preliminary separation step was not used before the application of parallel factor analysis Tucker3, three-way partial least squares and unfolded partial least squares approaches for the analysis of the related drug substances in samples. Three-way excitation-emission matrix data array was obtained by concatenating excitation-emission matrices of the calibration set, validation set, and commercial tablet samples. The excitation-emission matrix data array was used to get parallel factor analysis, Tucker3, three-way partial least squares and unfolded partial least squares calibrations and to predict the amounts of valsartan and amlodipine besylate in samples. For all the methods, calibration and prediction of valsartan and amlodipine besylate were performed in the working concentration ranges of 0.25–4.50μg/mL. The validity and the performance of all the proposed methods were checked by using the validation parameters. From the analysis results, it was concluded that the described two-way and three-way algorithmic methods were very useful for the simultaneous quantitative resolution and routine analysis of the related drug substances in marketed samples.

Identification of representative samples from existing samples for digital soil mapping

Existing sample data are important for digital soil mapping. Different sample points possess different representativeness. The representativeness of samples influences the soil mapping result greatly. However, few study focus on assessing the representativeness of single sample. In this paper, we proposed a method to identify representative samples from existing samples collected from multiple resources. The basic idea of the method was to use clusters of environmental covariates to approximate types of soil variations, and check the occupancy of the existing samples in centroids of environmental clusters. Those samples locating in typical locations or centroids of environmental clusters were considered as representative samples. In this paper, the proposed method was used to discern representative samples in 282 soil samples in Anhui Province, China. SOM content was mapped using a similarity based mapping method. Two cases with different training samples (representative samples, non-representative samples, and training samples including representative and non-representative samples) and validation samples were set to compare the mapping results and accuracies. The results showed that the SOM content maps predicted using representative training samples had generally higher accuracy than the results produced using non-representative samples, and comparative accuracies with the results produced using full training samples. To discern representative samples is helpful for understanding the soil-landscape relationships in an area and the proposed method can be used to design supplementary samples for a better soil mapping result. Mapping results and accuracies showed that different training and validation sample sets impacted the mapping results and accuracies greatly, which indicates that researchers should be cautious when using randomization to obtain training and validation subsets for soil mapping.

The first all-season sample set for mapping global land cover with Landsat-8 data

We report the world’s first all-season training and validation sample sets for global land cover classification with Landsat-8 data. Prior to this, such samples were only available at a single date primarily from the growing season. It is unknown how much limitation such a single-date sample has to mapping global land cover in other seasons of the year. To answer this question, we selected available Landsat-8 images from four seasons and collected training and validation samples from them. We compared the performances of training samples in different seasons using Random Forest algorithm. We found that the use of training samples from any individual season would result in the best overall classification accuracy when validated by samples in the same season. The global overall accuracy from combined best seasonal results was 67.2% when classifying the 11 Level-1 classes in the Finer Resolution Observation and Monitoring of Global Land Cover (FROM-GLC) classification system. The use of training samples from all seasons (named all-season training sample set hereafter) produced an overall accuracy of 67.0%. We also tested classification within 10° latitude 60° longitude zones using all-season training subsample within each zone and obtained an overall accuracy of 70.2%. This indicates that properly grouped subsamples in space can help improve classification accuracies. All the results in this study seem to suggest that it is possible to use an all-season training sample set to reach global optimality with universal applicability in classifying images acquired at any time of a year for global land cover mapping.

Development of near-infrared spectroscopy models for quantitative determination of cellulose and hemicellulose contents of big bluestem

Big bluestem is a dominant warm-season perennial native grass that has underutilized potential as a bioenergy crop. The objective of this study was to leverage a high-throughput, cost-effective phenotype of cellulose and hemicellulose contents in big bluestem biomass using near-infrared (NIR) spectroscopy to facilitate plant breeding and genetics studies. In order to develop NIR prediction models, a set of 56 big bluestem samples with seven genotypes from four planting locations in 2010 and 2011 were analyzed according to traditional wet chemical methods. Advanced multivariate analysis techniques and NIR spectroscopy improved the prediction models based on value of the coefficient of determination (R2). Partial least squares proved to be a better quantitative method than principal component regression based on larger R2, ratio of standard error of prediction set to sample standard deviation (RPD), and root mean square error of prediction (RMSEP) when developing NIR prediction models. The spectral range from 4000 to 7500 cm−1 with the first derivative treatment yielded a better prediction model than full range, with R2 of 0.92, RMSEP of 0.67%, and RPD of 4.52 in the validation sample set for cellulose and R2 of 0.91, RMSEP of 0.72%, and RPD of 3.12 for hemicellulose. These models provide good insight into the relationship between chemical bonds and structure sugars of big bluestem, allowing a rapid and accurate determination of cellulose and hemicellulose contents at low cost.