Principal Component Analysis Space Research Articles

Macrophage proteomic analysis of covalent immobilization of hyaluronic acid and graphene oxide on CoCr alloy in a tribocorrosive environment.

In this work, a sequential covalent immobilization of graphene oxide (GO) and hyaluronic acid (HA) is performed to obtain a biocompatible wear-resistant nanocoating on the surface of the biomedical grade cobalt-chrome (CoCr) alloy. Nanocoated CoCr surfaces were characterized by Raman spectroscopy and electrochemical impedance spectroscopy (EIS) in 3 g/L HA electrolyte. Tribocorrosion tests of the nanocoated CoCr surfaces were carried out in a pin on flat tribometer. The biological response of covalently HA/GO biofunctionalized CoCr surfaces with and without wear-corrosion processes was studied through the analysis of the proteome of macrophages. Raman spectra revealed characteristic bands of GO and HA on the functionalized CoCr surfaces. The electrochemical response by EIS showed a stable and protective behavior over 23 days in the simulated biological environment. HA/GO covalently immobilized on CoCr alloy is able to protect the surface and reduce the wear volume released under tribocorrosion tests. Unsupervised classification analysis of the macrophage proteome via hierarchical clustering and principal component analysis (PCA) revealed that the covalent functionalization on CoCr enhances the macrophage biocompatibility in vitro. On the other hand, disruption of the HA/GO nanocoating by tribocorrosion processes induced a macrophage proteome which was differently clustered and was distantly located in the PCA space. In addition, tribocorrosion induced an increase in the percentage of upregulated and downregulated proteins in the macrophage proteome, revealing that disruption of the covalent nanocoating impacts the macrophage proteome. Although macrophage inflammation induced by tribocorrosion of HA/GO/CoCr surfaces is observed, it is ameliorated by the covalently grafting of HA, which provides immunomodulation by eliciting downregulations in characteristic pro-inflammatory signaling involved in inflammation and aseptic loosening of CoCr joint arthroplasties. Covalent HA/GO nanocoating on CoCr provides potential applications for in vivo joint prostheses led a reduced metal-induced inflammation and degradation by wear-corrosion.

Quantitative Analysis of Bone Marrow Features Highlights Heterogeneity in Myelofibrosis Patients Treated with Zinpentraxin Alfa in a Phase II Clinical Study

Quantitative image analysis has potential to transform the interpretation of bone marrow trephine (BMT) samples in myeloproliferative neoplasms (MPN) and improve the evaluation of anti-fibrotic therapies. To investigate the utility of recently developed algorithms evaluating reticulin fibrosis and megakaryocyte features in myelofibrosis (MF), we analysed samples from a multi-center, phase II study of zinpentraxin alfa (ZPN, PRM-151). ZPN is a recombinant form of human pentraxin-2 (PTX2) that has shown clinical activity as monotherapy and in combination with ruxolitinib (RUX) in a phase II trial in patients with Int-1/-2 or high risk MF (NCT01981850). In stage 2 of this study, patients ineligible for, intolerant of, or with an inadequate response to RUX were randomised to receive 0.3, 3.0, or 10.0 mg/kg ZPN on Days 1, 3, and 5 of Cycle 1, and every 4 weeks thereafter for up to 9 cycles. We demonstrate the potential of quantitative image analysis to augment and refine conventional expert histological assessment of MF. Reticulin and H+E-stained BMTs from three timepoints (screening, C4D1 and C9D29) were analysed from 50/97 patients enrolled. Manual assessment (MA) of marrow fibrosis and megakaryocyte features was performed by blinded, independent central review. Assessment of fibrosis using the Continuous Indexing of Fibrosis (CIF) score and analysis of megakaryocytes was performed by automated analyses ( Ryou H, Leukaemia, 2022). There was a moderate correlation between the automated average CIF score and manual fibrosis grade (Spearman's rho = 0.39). However, there was significant overlap in the distribution of CIF scores between samples manually assigned to MF-2 and MF-3. Approximately 38% of MF-2 samples fell within the interquartile range of CIF distribution observed in MF-3, and around 48% of MF-3 samples fell within the interquartile range observed in MF-2. Next, we visualised the screening BMTs in fibrosis PCA (principal component analysis) space to capture both average CIF score and Shannon entropy (‘unevenness‘) of CIF score for each sample. This revealed significant heterogeneity in reticulin fibrosis at screening when compared to an independent reference cohort of 130 MPN samples ( Ryou H, Leukaemia, 2022), likely reflecting the effects of pre-treatment [ Figure 1]. To track fibrosis across all timepoints we plotted the average CIF score for each sample. This revealed striking variability in fibrotic response within the trial cohort, with 15 of 42 patients (36%) demonstrating improvement in average CIF score [ Figure 2]. Notably, improvements in average CIF score by cycle 9 appeared to be most marked in patients with higher CIF scores at screening. No significant ZPN dose-dependent effect on fibrosis was observed. We next measured megakaryocyte density (megakaryocytes per unit area of intertrabecular space) and compared it to MA. We found a strong correlation (Spearman's rho = 0.81) between automated and MA of megakaryocyte density, with no significant change across any of the treatment arms. For megakaryocyte clustering (defined as ≥ 3 megakaryocytes in direct contact) we observed a moderate to strong correlation (Spearman's rho = 0.62) between MA and automated assessment, with no significant change across any of the treatment arms. Finally, we sought to compare the sample megakaryocyte cell features (cytomorphology and topology) at screening with those of an independent reference cohort of 88 MPN samples ( Sirinukunwattana, Blood Adv., 2020). Strikingly, not only was there marked heterogeneity in megakaryocyte features between patient samples at screening, but the observed features were significantly different to those of newly diagnosed MF. No significant changes in megakaryocyte cytomorphology or topology were seen across any of the treatment arms. In summary, our findings reveal considerable BMT morphological heterogeneity in patients participating in a Phase II trial of ZPN. This likely reflects poorly understood and under-recognised variation in morphological features encountered in patients with longstanding and/or pre-treated MF when compared to newly diagnosed patients. This highlights the potential of such variability to confound the evaluation of novel therapeutics in MPN, and emphasises the utility of robust quantitative methods to analyse and visualise morphological features in clinical study samples that can complement conventional manual assessment.

BinderSpace: A package for sequence space analyses for datasets of affinity-selected oligonucleotides and peptide-based molecules.

Discovery of target-binding molecules, such as aptamers and peptides, is usually performed with the use of high-throughput experimental screening methods. These methods typically generate large datasets of sequences of target-binding molecules, which can be enriched with high affinity binders. However, the identification of the highest affinity binders from these large datasets often requires additional low-throughput experiments or other approaches. Bioinformatics-based analyses could be helpful to better understand these large datasets and identify the parts of the sequence space enriched with high affinity binders. BinderSpace is an open-source Python package that performs motif analysis, sequence space visualization, clustering analyses, and sequence extraction from clusters of interest. The motif analysis, resulting in text-based and visual output of motifs, can also provide heat maps of previously measured user-defined functional properties for all the motif-containing molecules. Users can also run principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE) analyses on whole datasets and on motif-related subsets of the data. Functionally important sequences can also be highlighted in the resulting PCA and t-SNE maps. If points (sequences) in two-dimensional maps in PCA or t-SNE space form clusters, users can perform clustering analyses on their data, and extract sequences from clusters of interest. We demonstrate the use of BinderSpace on a dataset of oligonucleotides binding to single-wall carbon nanotubes in the presence and absence of a bioanalyte, and on a dataset of cyclic peptidomimetics binding to bovine carbonic anhydrase protein. BinderSpace is openly accessible to the public via the GitHub website: https://github.com/vukoviclab/BinderSpace.

Phenolic compounds, methylxanthines, and preference drivers of dark chocolate made with hybrid cocoa beans

The distinct properties of cocoa beans may be due to their different geographical and genetic origins which lead to chocolates with different characteristics. This study aimed to evaluate dark chocolate samples made with hybrid cocoa cultivated in the Bahia State, Brazil, regarding the composition of bioactive compounds (spectrophotometry, High Performance Liquid Chromatography) and the sensory characteristics (Quantitative Descriptive Analysis-QDA®), and acceptance test). External preference mapping was conducted by first using principal components analysis (PCA) on the QDA data, and then relating consumer preference (overall liking) to this PCA space by regression analysis. Significant differences among the cocoa varieties were found. Chocolate samples that showed lower acceptability also presented higher content of theobromine, caffeine, anthocyanins, and gallic acid which conferred greater intensities of bitterness, astringency, and intense cocoa flavor. Lower levels of methylxanthines, monomeric phenolic compounds, and anthocyanins allowed the highlight of fruity, sweet and caramelized notes, which contributed to greater acceptance of the chocolate samples. Therefore, these characteristics can be recognized as consumer preference drivers (negative and positive, respectively) for the high cocoa chocolates. It seems that higher values of ephicatechin and catechin had no negative contribution to the flavor of the chocolate samples. The findings of this study can be used in agriculture and in the cocoa/chocolate industry, suggesting the potential and applicability of promising hybrid cocoa, with important characteristics, representing a genetic material of great quality, that may be classified as fine flavor cocoa, associating greater value to the product, and contributing to meet consumer needs.

Demonstration of a Principal Component Analysis Trajectory Method to Assess Bioremediation Progress at a TCE‐Impacted Site

AbstractIn‐situ bioremediation (ISB) is a popular remediation technology for the treatment of a range of compounds, including chlorinated solvents such as tetrachloroethene and trichloroethene (TCE). Large amounts of data are collected before, during, and after ISB applications to determine amendment approaches, monitor progress and evaluate success. The interpretation of these large datasets can be limited by the tools and techniques used for data analysis, and there is considerable potential in applying data reduction and multivariate techniques used elsewhere to performance monitoring during ISB. In this study, a principal component analysis (PCA) trajectory method was applied to a TCE‐impacted ISB site dataset, as an alternative to the inspection of time series data. The method connected each monitoring well's scores through PCA space to account for temporal changes in multiple analytes across the site. The method was used to separate monitoring well locations into categories that included On‐track and Unsuccessful based on their similarity to background wells in PCA space. The results agreed with those generated using traditional methods (e.g., time series plots) and were able to efficiently summarize large amounts of data to facilitate interpretation. It is expected that this PCA trajectory method could provide a useful screening tool to quickly identify site‐wide trends for the evaluation of ISB performance.

Material parameter identification of the proximal and distal segments of the porcine thoracic aorta based on ECG-gated CT angiography

Vessel wall material parameters are important in biomechanical research. The purpose of this study was to identify the material parameters of two porcine thoracic aortic segments and verify the accuracy of the identification results with uniaxial tensile testing. Principal component analysis (PCA) was used to reduce the dimensionality of the stress matrix. Data points in PCA space were initially screened by K-means cluster analysis, and connection networks of two levels were constructed based on the distance between the data points. The material parameters corresponding to the data points were substituted, and pressure was applied to convert the diastolic models to systolic models to match those reconstructed from electrocardiographic (ECG) gated computed tomography angiography (CTA), and determine the optimal material parameters. The proximal and distal segments of the thoracic aorta were selected for uniaxial tensile testing, and the stress–strain curves obtained from numerical simulations and experiments were compared. The average distances between the simulated systolic proximal and distal segments and their corresponding systolic models reconstructed from the ECG-gated-CTA were 0.388 mm and 0.257 mm, respectively. The fit goodness of the stress–strain data obtained by two methods was 0.9953 and 0.9750, respectively, with equivalent elastic moduli differences of 1.08% and 0.36%. Thus, a material parameter screening method for different aortic segments was proposed and its accuracy was verified experimentally with good consistency. This method is expected to provide a theoretical basis for biomechanical studies of aortic diseases.

A geometric relationship of F2, F3 and F4-statistics with principal component analysis.

Principal component analysis (PCA) and F-statistics sensu Patterson are two of the most widely used population genetic tools to study human genetic variation. Here, I derive explicit connections between the two approaches and show that these two methods are closely related. F-statistics have a simple geometrical interpretation in the context of PCA, and orthogonal projections are a key concept to establish this link. I show that for any pair of populations, any population that is admixed as determined by an F3-statistic will lie inside a circle on a PCA plot. Furthermore, the F4-statistic is closely related to an angle measurement, and will be zero if the differences between pairs of populations intersect at a right angle in PCA space. I illustrate my results on two examples, one of Western Eurasian, and one of global human diversity. In both examples, I find that the first few PCs are sufficient to approximate most F-statistics, and that PCA plots are effective at predicting F-statistics. Thus, while F-statistics are commonly understood in terms of discrete populations, the geometric perspective illustrates that they can be viewed in a framework of populations that vary in a more continuous manner.This article is part of the theme issue ‘Celebrating 50 years since Lewontin's apportionment of human diversity’.

Analysis of size and concentration of microplastics in water using static light scattering combined with PCA and LDA

Quantitative analysis of size and concentration of microplastics is a crucial step for having a better understanding of plastic pollution in the environment. Such information is typically obtained in a single particle mode that significantly increases the analysis time and can be a cumbersome task. Therefore, we demonstrate here a measurement technique based on Static Light Scattering (SLS) combined with chemometric methods such as Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) for resolving the size and concentration of multiple microplastic particles in water. Two sets of samples with uniform and non-uniform size distribution of microplastics, called “monodisperse” and “polydisperse”, respectively, are fully investigated. It is shown that a relationship exists between the scattering signals of mono- and polydisperse samples on the PCA space. Hence, a PCA-LDA model that is constructed on the PCA space of monodisperse samples is used to discriminate the size of the microplastics in polydisperse samples. By specifying the size of the particles, their concentration is determined using a simple linear fit.

Discrimination of waterborne pathogens, Cryptosporidium parvum oocysts and bacteria using surface-enhanced Raman spectroscopy coupled with principal component analysis and hierarchical clustering

Waterborne pathogens (parasites, bacteria) are serious threats to human health. Cryptosporidium parvum is one of the protozoan parasites that can contaminate drinking water and lead to diarrhea in animals and humans. Rapid and reliable detection of these kinds of waterborne pathogens is highly essential. Yet, current detection techniques are limited for waterborne pathogens and time-consuming and have some major drawbacks. Therefore, rapid screening methods would play an important role in controlling the outbreaks of these pathogens. Here, we used label-free surface-enhanced Raman Spectroscopy (SERS) combined with multivariate analysis for the detection of C. parvum oocysts along with bacterial contaminants including, Escherichia coli, and Staphylococcus aureus. Silver nanoparticles (AgNPs) are used as SERS substrate and samples were prepared with simply mixed of concentrated AgNPs with microorganisms. Each species presented distinct SERS spectra. Principal component analysis (PCA) and hierarchical clustering were performed to discriminate C. parvum oocysts, E. coli, and S. aureus. PCA was used to visualize the dataset and extract significant spectral features. According to score plots in 3 dimensional PCA space, species formed distinct group. Furthermore, each species formed different clusters in hierarchical clustering. Our study indicates that SERS combined with multivariate analysis techniques can be utilized for the detection of C. parvum oocysts quickly.

A principal component analysis (PCA) based assessment of the gait performance.

The gait assessment is instrumental for evaluating the efficiency of rehabilitation of persons with a motor impairment of the lower extremities. The protocol for quantifying the gait performance needs to be simple and easy to implement; therefore, a wearable system and user-friendly computer program are preferable. We used the Gait Master (instrumented insoles) with the industrial quality ground reaction forces (GRF) sensors and 6D inertial measurement units (IMU). WiFi transmitted 10 signals from the GRF sensors and 12 signals from the accelerometers and gyroscopes to the host computer. The clinician was following in real-time the acquired data to be assured that the WiFi operated correctly. We developed a method that uses principal component analysis (PCA) to provide a clinician with easy to interpret cyclograms showing the difference between the recorded and healthy-like gait performance. The cyclograms formed by the first two principal components in the PCA space show the step-to-step reproducibility. We suggest that a cyclogram and its orientation to the coordinate system PC1 vs. PC2 allow a simple assessment of the gait. We show results for six healthy persons and five patients with hemiplegia.

Joint analysis of structural connectivity and cortical surface features: correlates with mild traumatic brain injury.

Mild traumatic brain injury (mTBI) is a complex syndrome that affects up to 600 per 100,000 individuals, with a particular concentration among military personnel. About half of all mTBI patients experience a diverse array of chronic symptoms which persist long after the acute injury. Hence, there is an urgent need for better understanding of the white matter and gray matter pathologies associated with mTBI to map which specific brain systems are impacted and identify courses of intervention. Previous works have linked mTBI to disruptions in white matter pathways and cortical surface abnormalities. Herein, we examine these hypothesized links in an exploratory study of joint structural connectivity and cortical surface changes associated with mTBI and its chronic symptoms. Briefly, we consider a cohort of 12 mTBI and 26 control subjects. A set of 588 cortical surface metrics and 4,753 structural connectivity metrics were extracted from cortical surface regions and diffusion weighted magnetic resonance imaging in each subject. Principal component analysis (PCA) was used to reduce the dimensionality of each metric set. We then applied independent component analysis (ICA) both to each PCA space individually and together in a joint ICA approach. We identified a stable independent component across the connectivity-only and joint ICAs which presented significant group differences in subject loadings (p<0.05, corrected). Additionally, we found that two mTBI symptoms, slowed thinking and forgetfulness, were significantly correlated (p<0.05, corrected) with mTBI subject loadings in a surface-only ICA. These surface-only loadings captured an increase in bilateral cortical thickness.

GC-MS and HPLC-DAD analysis of fatty acid profile and functional phytochemicals in fifty cold-pressed plant oils in Thailand

Cold-pressed oil is one of the healthiest plant extracts, but its use is limited only in some kinds of plants. Therefore, we aimed to investigate some potential cold-pressed oils with attractive fatty acid profiles and high amounts of functional phytochemicals. Fifty cold-pressed plant oils were prepared from various plant materials in Thailand, in which some of them were from uncommon or unattended plant materials. The oils included were nut oils (n = 9), pseudo-cereal oils (n = 9), legume oils (n = 3), amaranth oils (n = 3), marrow seed oils (n = 8), cruciferous seed oils (n = 7), and leafy green seed oils (n = 11). Gas-chromatography mass-spectrometry (GC-MS) and high-performance liquid chromatography coupled with a diode array detector (HPLC-DAD) were employed to analyze fatty acid profile and five functional phytochemicals (e.g., phytosterols, cholecalciferol, and squalene). Saturated fatty acids were detected around 7.87–36.04%, monounsaturated fatty acids 10.17–80.25%, and polyunsaturated fatty acids nondetectable (ND)–78.25%, phytosterols 663–15123 μg g−1, squalene 265–5979 μg g−1, and cholecalciferol ND–1287.75 μg g−1. The study showed chemical characteristic of the analyzed oils: some contained good fatty acid composition and some were rich in functional phytochemical content. Among the obtained oils, marrow seed oils are a good source of phytosterol, cholecalciferol, and linoleic acid. Pseudo-cereal oils are rich in squalene and linolenic acid. Legume oils are rich in phytosterols and oleic acid. Besides, principal component analysis (PCA) was applied to identify the significance of oils that share compositional similarity (e.g., the samples from pseudo-cereal oil were found on the lower side of the PCA space, which separated them from marrow and leafy green seed oils distributed on the upper part of the plot). In summary, the qualitative and quantitative data would provide a good foundation for further application or selection of those plant oils for health purposes.

Improved prediction of Canada lynx distribution through regional model transferability and data efficiency.

The application of species distribution models (SDMs) to areas outside of where a model was created allows informed decisions across large spatial scales, yet transferability remains a challenge in ecological modeling. We examined how regional variation in animal‐environment relationships influenced model transferability for Canada lynx (Lynx canadensis), with an additional conservation aim of modeling lynx habitat across the northwestern United States. Simultaneously, we explored the effect of sample size from GPS data on SDM model performance and transferability. We used data from three geographically distinct Canada lynx populations in Washington (n = 17 individuals), Montana (n = 66), and Wyoming (n = 10) from 1996 to 2015. We assessed regional variation in lynx‐environment relationships between these three populations using principal components analysis (PCA). We used ensemble modeling to develop SDMs for each population and all populations combined and assessed model prediction and transferability for each model scenario using withheld data and an extensive independent dataset (n = 650). Finally, we examined GPS data efficiency by testing models created with sample sizes of 5%–100% of the original datasets. PCA results indicated some differences in environmental characteristics between populations; models created from individual populations showed differential transferability based on the populations' similarity in PCA space. Despite population differences, a single model created from all populations performed as well, or better, than each individual population. Model performance was mostly insensitive to GPS sample size, with a plateau in predictive ability reached at ~30% of the total GPS dataset when initial sample size was large. Based on these results, we generated well‐validated spatial predictions of Canada lynx distribution across a large portion of the species' southern range, with precipitation and temperature the primary environmental predictors in the model. We also demonstrated substantial redundancy in our large GPS dataset, with predictive performance insensitive to sample sizes above 30% of the original.

Prediction of Relatively High-Energy Seismic Events Using Spatial–Temporal Parametrisation of Mining-Induced Seismicity

Mining-induced seismicity has been reported almost in every mining country. Large seismic events with high-energy (HE) radiation can pose a serious threat to safe mining operations, which are the direct cause for rockbursts in underground mines. Over the last 30 years, statistical techniques to parameterise seismic data and predict seismic hazard have been developed significantly with promising results. However, similar to earthquake prediction, the prediction accuracy of HE seismic events remains a challenging task due to the complex nature of mining-induced seismic events. This paper aims to parameterise spatial, temporal and energy information conveyed by past seismic activities to provide early warnings for HE events. The 4D spatial–temporal seismic data were transferred to a 2D plane using principal component analysis (PCA) and then applied with a kernel density estimator to calculate their probability density function. The mode of probability density distribution, ρ(x)max, was proposed as a measure to quantify the clustering degree of past seismic events in the PCA space. The peaks of ρ(x)max were found as an effective precursor to predicting the onset time of future HE events. After inverse transformation, the spatial locations of ρ(x)max also indicate the potential high-risk areas that susceptible to HE events, which can be used to predict the onset location of HE events. As a supplement to the spatial–temporal information, energy/magnitude information was parameterised by the modified cumulative Benioff strain and b value. The accuracy of using these seismic parameters for the prediction of HE events has been assessed based on the seismic data collected from a Chinese coal mine.

Spiking neural network approaches PCA with metaheuristics

This Letter presents meaningful results that demonstrate the reduction of dimensionality by spiking neural networks (SNNs) on benchmarking data. This experimental scheme includes metaheuristics, namely, the artificial bee colony algorithm (ABC algorithm) for finding optimal conductance values in the SNNs. Therefore, the objective function in the used ABC algorithm leads the SNNs to compute the principal component analysis (PCA), efficiently. The eigendecomposition of the information drawn by the SNNs in the training phase is the base of the formulated objective function. In these experiments, the Izhikevich model represents the spiking neurons, which have biological plausibility with parameters for reproducing a uniform firing rate. The visualisation of clusters in the 3D PCA space, whose sample values are compared with the PCA function in Matlab, is also shown; this comparison demonstrates an acceptable error in the MSE sense.

Optimizing exoplanet atmosphere retrieval using unsupervised machine-learning classification

ABSTRACT One of the principal bottlenecks to atmosphere characterization in the era of all-sky surveys is the availability of fast, autonomous, and robust atmospheric retrieval methods. We present a new approach using unsupervised machine learning to generate informed priors for retrieval of exoplanetary atmosphere parameters from transmission spectra. We use principal component analysis (PCA) to efficiently compress the information content of a library of transmission spectra forward models generated using the platon package. We then apply a k-means clustering algorithm in PCA space to segregate the library into discrete classes. We show that our classifier is almost always able to instantaneously place a previously unseen spectrum into the correct class, for low-to-moderate spectral resolutions, R, in the range R = 30−300 and noise levels up to 10 per cent of the peak-to-trough spectrum amplitude. The distribution of physical parameters for all members of the class therefore provides an informed prior for standard retrieval methods such as nested sampling. We benchmark our informed-prior approach against a standard uniform-prior nested sampler, finding that our approach is up to a factor of 2 faster, with negligible reduction in accuracy. We demonstrate the application of this method to existing and near-future observatories, and show that it is suitable for real-world application. Our general approach is not specific to transmission spectroscopy and should be more widely applicable to cases that involve the repetitive fitting of trusted high-dimensional models to large data catalogues, including beyond exoplanetary science.

An NMR-Based Similarity Metric for Higher Order Structure Quality Assessment Among U.S. Marketed Insulin Therapeutics

Protein or peptide higher order structure (HOS) is a quality attribute that could affect therapeutic efficacy and safety. Where appropriate, the HOS similarity between a proposed follow-on product and the reference listed drug should be demonstrated during regulatory assessment. Establishing quantitative HOS similarity for 2 drug substances, manufactured by different processes, has been challenging. Herein, HOS differences among U.S. marketed insulin drug products (DPs) were quantified using nuclear magnetic resonance spectra and principal component analysis (PCA). Then, the unitless Mahalanobis distance (DM) in PCA space was calculated between insulin analog reference listed drugs and their recently approved follow-on products, and all DM values were 3.29 or less. By contrast, a larger DM value of 20.5 was obtained between the 2 insulin human DPs independently approved. However, upon mass-balanced and reversible dialysis of the 2 insulin human DPs against the same buffers, the DM value was reduced to 1.19 or less. Thus, the observed range of nuclear magnetic resonance-PCA–derived DM values can be used as a robust and sensitive measure of HOS similarity. Overall, the DM values of 3.3 for DP and 1.2 for drug substances using insulin therapeutics represented realistic and achievable similarity metrics for developing generic or biosimilar drugs, quality assurance, or control.

S16-08 SESSION 16: PLANNING/IMAGING – PART II STATISTICAL SHAPE ANALYSIS OF FETAL FACE MORPHOLOGY FROM 3D ULTRASOUND IMAGES

Introduction: Antenatal detection of facial dysmorphism using 3D ultrasound may raise suspicion of underlying genetic conditions, but infrequently leads to definitive antenatal diagnosis. Despite advances in array and non-invasive prenatal testing, not all genetic conditions can be ascertained from such testing. The aim of this study was to investigate the feasibility of quantitative assessment of fetal face features using prenatal 3D ultrasound volumes and statistical shape modelling, to objectively distinguish dysmorphic from normal cases. Methods: 13 normal and 7 abnormal stored 3D ultrasound (Voluson S8-E8, GE Healthcare) fetal face volumes were analysed, at a median gestation of 29 + 4 weeks [25 + 0 – 36 + 1]. The 20 volumes were post-processed to generate 3D surface meshes, subsequently aligned. The surfaces served as input for statistical shape modelling (Deformetrica) to compute the population mean 3D face shape. Principal component analysis (PCA), a mathematical technique used to condense a large amount of 3D data into a smaller number of relevant variables (shape modes), was used to study the 3D shape variations in the population. Results: 10 PCA modes explained >90% of the total shape variability in the population. While the first mode accounted for overall size differences, the second highlighted shape feature changes from an overall proportionate towards a more asymmetric face shape, with wide prominent forehead and undersized, posteriorly positioned chin. Analysis of the subject specific Mahalanobis distance in PCA space suggested differences between normal and abnormal fetuses (median and interquartile range distance values 7.3±5.6 for the normal group vs. 13.3±9.8 for the abnormal group, p=.056). Conclusion: In conclusion, this feasibility study demonstrates that objective characterisation and quantification of fetal facial morphology is possible from 3D ultrasound. This technique has the potential to assist in-utero diagnosis, particularly of rare conditions where facial dysmorphology is a feature.

Filling missing values by local reconstruction for incomplete label distribution learning

Label distribution learning (LDL) deals with the problems when we care more about the relative importance of different labels in the description of an instance, where labels are associated with each instance to some degree. LDL has achieved great success in many applications, but most existing LDL methods cannot handle learning tasks with incomplete annotation information. In this paper, we propose a novel incomplete label distribution learning method based on local reconstruction (IncomLDL-LR). Both the feature matrix and label information are included in a unified Principal Component Analysis (PCA) model to describe the intrinsic structure of original data in the supervised low-dimensional space. Based on the reasonable assumption that the incomplete label of each instance can be linearly reconstructed from its neighbours' labels, IncomLDL-LR gradually recovers the missing label values by the averaged column score of corresponding neighbours in the PCA space. The proposed algorithms are compared with state-of-the-art algorithms using five LDL evaluation metrics on 15 public datasets. Extensive experiments validate the effectiveness of our proposal.

Filling missing values by local reconstruction for incomplete label distribution learning

Label distribution learning (LDL) deals with the problems when we care more about the relative importance of different labels in the description of an instance, where labels are associated with each instance to some degree. LDL has achieved great success in many applications, but most existing LDL methods cannot handle learning tasks with incomplete annotation information. In this paper, we propose a novel incomplete label distribution learning method based on local reconstruction (IncomLDL-LR). Both the feature matrix and label information are included in a unified Principal Component Analysis (PCA) model to describe the intrinsic structure of original data in the supervised low-dimensional space. Based on the reasonable assumption that the incomplete label of each instance can be linearly reconstructed from its neighbours' labels, IncomLDL-LR gradually recovers the missing label values by the averaged column score of corresponding neighbours in the PCA space. The proposed algorithms are compared with state-of-the-art algorithms using five LDL evaluation metrics on 15 public datasets. Extensive experiments validate the effectiveness of our proposal.