Machine Learning in Healthcare

Recent advancements in Artificial Intelligence (AI) and Machine Learning (ML) technology have brought on substantial strides in predicting and identifying health emergencies, disease populations, and disease state and immune response, amongst a few. Although, skepticism remains regarding the practical application and interpretation of results from ML-based approaches in healthcare settings, the inclusion of these approaches is increasing at a rapid pace. Here we provide a brief overview of machine learning-based approaches and learning algorithms including supervised, unsupervised, and reinforcement learning along with examples. Second, we discuss the application of ML in several healthcare fields, including radiology, genetics, electronic health records, and neuroimaging. We also briefly discuss the risks and challenges of ML application to healthcare such as system privacy and ethical concerns and provide suggestions for future applications.     

Pitfalls in TCR gene clonality testing: teaching cases

Clonality testing in T-lymphoproliferations has technically become relatively easy to perform in routine laboratories using standardized multiplex polymerase chain reaction protocols for T-cell receptor (TCR) gene analysis as developed by the BIOMED-2 Concerted Action BMH4-CT98-3936. Expertise with clonality diagnostics and knowledge about the biology of TCR gene recombination are essential for correct interpretation of TCR clonality data. Several immunobiological and technical pitfalls that should be taken into account to avoid misinterpretation of data are addressed in this report. Furthermore, we discuss the need to integrate the molecular data with those from immunohistology, and preferably also flow cytometric immunophenotyping, for appropriate interpretation. Such an interactive, multidisciplinary diagnostic model guarantees integration of available data to reach the most reliable diagnosis.     

Network-assisted analysis to prioritize GWAS results: principles, methods and perspectives

Genome-wide association studies (GWAS) have rapidly become a powerful tool in genetic studies of complex diseases and traits. Traditionally, single marker-based tests have been used prevalently in GWAS and have uncovered tens of thousands of disease-associated SNPs. Network-assisted analysis (NAA) of GWAS data is an emerging area in which network-related approaches are developed and utilized to perform advanced analyses of GWAS data in order to study various human diseases or traits. Progress has been made in both methodology development and applications of NAA in GWAS data, and it has already been demonstrated that NAA results may enhance our interpretation and prioritization of candidate genes and markers. Inspired by the strong interest in and high demand for advanced GWAS data analysis, in this review article, we discuss the methodologies and strategies that have been reported for the NAA of GWAS data. Many NAA approaches search for subnetworks and assess the combined effects of multiple genes participating in the resultant subnetworks through a gene set analysis. With no restriction to pre-defined canonical pathways, NAA has the advantage of defining subnetworks with the guidance of the GWAS data under investigation. In addition, some NAA methods prioritize genes from GWAS data based on their interconnections in the reference network. Here, we summarize NAA applications to various diseases and discuss the available options and potential caveats related to their practical usage. Additionally, we provide perspectives regarding this rapidly growing research area.     

Systemic Perturbation of the ERK Signaling Pathway by the Proteasome Inhibitor,MG132

Inhibition of the ubiquitin-proteasome protein degradation pathway has been identified as a viable strategy for anti-tumor therapy based on its broad effects on cell proliferation. By the same token, the variety of elicited effects confounds the interpretation of cell-based experiments using proteasome inhibitors such as MG132. It has been proposed that MG132 treatment reduces growth factor-stimulated phosphorylation of extracellular signal-regulated kinases (ERKs), at least in part through upregulation of dual specificity phosphatases (DUSPs). Here, we show that the effects of MG132 treatment on ERK signaling are more widespread, leading to a reduction in activation of the upstream kinase MEK. This suggests that MG132 systemically perturbs the intracellular phosphoproteome, impacting ERK signaling by reducing phosphorylation status at multiple levels of the kinase cascade.     

BMP signaling in wing development: A critical perspective on quantitative image analysis

Bone Morphogenetic Proteins (BMPs) are critical for pattern formation in many animals. In numerous tissues, BMPs become distributed in spatially non-uniform profiles. The gradients of signaling activity can be detected by a number of biological assays involving fluorescence microscopy. Quantitative analyses of BMP gradients are powerful tools to investigate the regulation of BMP signaling pathways during development. These approaches rely heavily on images as spatial representations of BMP activity levels, using them to infer signaling distributions that inform on regulatory mechanisms. In this perspective, we discuss current imaging assays and normalization methods used to quantify BMP activity profiles with a focus on the Drosophila wing primordium. We find that normalization tends to lower the number of samples required to establish statistical significance between profiles in controls and experiments, but the increased resolvability comes with a cost. Each normalization strategy makes implicit assumptions about the biology that impacts our interpretation of the data. We examine the tradeoffs for normalizing versus not normalizing, and discuss their impacts on experimental design and the interpretation of resultant data.     

Nonredundant Regulation of Rice Arbuscular Mycorrhizal Symbiosis by Two Members of the PHOSPHATE TRANSPORTER1 Gene Family

Pi acquisition of crops via arbuscular mycorrhizal (AM) symbiosis is becoming increasingly important due to limited high-grade rock Pi reserves and a demand for environmentally sustainable agriculture. Here, we show that 70% of the overall Pi acquired by rice (Oryza sativa) is delivered via the symbiotic route. To better understand this pathway, we combined genetic, molecular, and physiological approaches to determine the specific functions of two symbiosis-specific members of the PHOSPHATE TRANSPORTER1 (PHT1) gene family from rice, ORYsa;PHT1;11 (PT11) and ORYsa;PHT1;13 (PT13). The PT11 lineage of proteins from mono- and dicotyledons is most closely related to homologs from the ancient moss, indicating an early evolutionary origin. By contrast, PT13 arose in the Poaceae, suggesting that grasses acquired a particular strategy for the acquisition of symbiotic Pi. Surprisingly, mutations in either PT11 or PT13 affected the development of the symbiosis, demonstrating that both genes are important for AM symbiosis. For symbiotic Pi uptake, however, only PT11 is necessary and sufficient. Consequently, our results demonstrate that mycorrhizal rice depends on the AM symbiosis to satisfy its Pi demands, which is mediated by a single functional Pi transporter, PT11.     

Visualizing density maps with UCSF Chimera

We describe methods for interactive visualization and analysis of density maps available in the UCSF Chimera molecular modeling package. The methods enable segmentation, fitting, coarse modeling, measuring and coloring of density maps for elucidating structures of large molecular assemblies such as virus particles, ribosomes, microtubules, and chromosomes. The methods are suitable for density maps with resolutions in the range spanned by electron microscope single particle reconstructions and tomography. All of the tools described are simple, robust and interactive, involving computations taking only seconds. An advantage of the UCSF Chimera package is its integration of a large collection of interactive methods. Interactive tools are sufficient for performing simple analyses and also serve to prepare input for and examine results from more complex, specialized, and algorithmic non-interactive analysis software. While both interactive and non-interactive analyses are useful, we discuss only interactive methods here.     

Genomewide rapid association using mixed model and regression: a fast and simple method for genomewide pedigree-based quantitative trait loci association analysis

For pedigree-based quantitative trait loci (QTL) association analysis, a range of methods utilizing within-family variation such as transmission-disequilibrium test (TDT)-based methods have been developed. In scenarios where stratification is not a concern, methods exploiting between-family variation in addition to within-family variation, such as the measured genotype (MG) approach, have greater power. Application of MG methods can be computationally demanding (especially for large pedigrees), making genomewide scans practically infeasible. Here we suggest a novel approach for genomewide pedigree-based quantitative trait loci (QTL) association analysis: genomewide rapid association using mixed model and regression (GRAMMAR). The method first obtains residuals adjusted for family effects and subsequently analyzes the association between these residuals and genetic polymorphisms using rapid least-squares methods. At the final step, the selected polymorphisms may be followed up with the full measured genotype (MG) analysis. In a simulation study, we compared type 1 error, power, and operational characteristics of the proposed method with those of MG and TDT-based approaches. For moderately heritable (30%) traits in human pedigrees the power of the GRAMMAR and the MG approaches is similar and is much higher than that of TDT-based approaches. When using tabulated thresholds, the proposed method is less powerful than MG for very high heritabilities and pedigrees including large sibships like those observed in livestock pedigrees. However, there is little or no difference in empirical power of MG and the proposed method. In any scenario, GRAMMAR is much faster than MG and enables rapid analysis of hundreds of thousands of markers.     

Mating Group Size and Stability in Reindeer Rangifer tarandus: The Effects of Male Characteristics,Sex Ratio and Male Age Structure

In polygynous mating systems, males compete intensely for mates and may mate several females during a single reproductive season. Accordingly, factors influencing the ability of males to control a larger number of females during the breeding season can provide information on the processes underlying sexual selection. In ungulates, age, body mass and social rank are considered good predictors of the reproductive success of males, but how male age structure and sex ratio in the population influence mating group (MG) dynamics has received little empirical testing. Between 1996 and 2005, we manipulated male age‐ and sex structure and monitored MG dynamics in a reindeer (Rangifer tarandus) population. We investigated the influence of male characteristics, percentage of males and male age structure on MG size and stability. We found that males with higher social rank (that were also older and heavier) controlled larger MGs (therefore had greater mating opportunities) and had more stable MGs (corresponding to a higher ability to maintain and control females) than males of lower social rank. Moreover, MG size and MG stability decreased as the percentage of males in the population increased, most likely resulting from greater male–male competition and increased female movements. Male age structure did not influence MG stability. Given the positive relationship between mating success and MG size (and likely MG stability), frequent female movements and intense competition among males to control females seem to be the principal components of reindeer MGs dynamic.     

Principles from community and metapopulation ecology: application to fungal entomopathogens

Fungal entomopathogens are often studied within the context of their use for biological control, yet these natural enemies are also excellent subjects for studies of ecological interactions. Here, we present selected principles from community ecology and discuss these in relation to fungal entomopathogens. We discuss the relevance of apparent competition, food web construction, intraguild predation and density-mediated and trait-mediated indirect effects. Although current knowledge of community interactions involving fungal entomopathogens are limited, fungal entomopathogens can be important, interactive members of communities and the activities of fungal entomopathogens should be evaluated in the context of ecological principles. We also discuss aspects of metapopulation ecology and the application of these principles to fungal entomopathogens. Knowledge of ecological interactions is crucial if we are to understand and predict the effects of fungal entomopathogens on host populations and understand the interactions among fungal entomopathogens and other organisms in the communities in which they occur.     

An alternative theoretical approach to escape decision-making: the role of visual cues

Escape enables prey to avoid an approaching predator. The escape decision-making process has traditionally been interpreted using theoretical models that consider ultimate explanations based on the cost/benefit paradigm. Ultimate approaches, however, suffer from inseparable extra-assumptions due to an inability to accurately parameterize the model's variables and their interactive relationships. In this study, we propose a mathematical model that uses intensity of predator-mediated visual stimuli as a basic cue for the escape response. We consider looming stimuli (i.e. expanding retinal image of the moving predator) as a cue to flight initiation distance (FID; distance at which escape begins) of incubating Mallards (Anas platyrhynchos). We then examine the relationship between FID, vegetation cover and directness of predator trajectory, and fit the resultant model to experimental data. As predicted by the model, vegetation concealment and directness of predator trajectory interact, with FID decreasing with increased concealment during a direct approach toward prey, but not during a tangential approach. Thus, we show that a simple proximate expectation, which involves only visual processing of a moving predator, may explain interactive effects of environmental and predator-induced variables on an escape response. We assume that our proximate approach, which offers a plausible and parsimonious explanation for variation in FID, may serve as an evolutionary background for traditional, ultimate explanations and should be incorporated into interpretation of escape behavior.     

New insights into butterfly-environment relationships using partitioning methods

Variation partitioning and hierarchical partitioning are novel statistical approaches that provide deeper understanding of the importance of different explanatory variables for biodiversity patterns than traditional regression methods. Using these methods, the variation in occupancy and abundance of the clouded apollo butterfly (Parnassius mnemosyne L.) was decomposed into independent and joint effects of larval and adult food resources, microclimate and habitat quantity. The independent effect of habitat quantity variables (habitat area and connectivity) captured the largest fraction of the variation in the clouded apollo patterns, but habitat connectivity had a major contribution only for occupancy data. The independent effects of resources and microclimate were higher on butterfly abundance than on occupancy. However, a considerable amount of variation in the butterfly patterns was accounted for by the joint effects of predictors and may thus be causally related to two or all three groups of variables. Abundance of the butterfly in the surroundings of the focal grid cell had a significant effect in all analyses, independently of the effects of other predictors. Our results encourage wider applications of partitioning methods in biodiversity studies.     

MG53 and disordered metabolism in striated muscle

MG53 is a member of tripartite motif family (TRIM) that expressed most abundantly in striated muscle. Using rodent models, many studies have demonstrated the MG53 not only facilitates membrane repair after ischemia reperfusion injury, but also contributes to the protective effects of both pre- and post-conditioning. Recently, however, it has been shown that MG53 participates in the regulation of many metabolic processes, especially insulin signaling pathway. Thus, sustained overexpression of MG53 may contribute to the development of various metabolic disorders in striated muscle. In this review, using cardiac muscle as an example, we will discuss muscle metabolic disturbances associated with diabetes and the current understanding of the underlying molecular mechanisms; in particular, the pathogenesis of diabetic cardiomyopathy. We will focus on the pathways that MG53 regulates and how the dysregulation of MG53 leads to metabolic disorders, thereby establishing a causal relationship between sustained upregulation of MG53 and the development of muscle insulin resistance and metabolic disorders. This article is part of a Special issue entitled Cardiac adaptations to obesity, diabetes and insulin resistance, edited by Professors Jan F.C. Glatz, Jason R.B. Dyck and Christine Des Rosiers.     

Super-resolution localization microscopy with photoactivatable fluorescent marker proteins

Fluorescent proteins (FPs) have become popular imaging tools because of their high specificity, minimal invasive labeling and allowing visualization of proteins and structures inside living organisms. FPs are genetically encoded and expressed in living cells, therefore, labeling involves minimal effort in comparison to approaches involving synthetic dyes. Photoactivatable FPs (paFPs) comprise a subclass of FPs that can change their absorption/emission properties such as brightness and color upon irradiation. This methodology has found a broad range of applications in the life sciences, especially in localization-based super-resolution microscopy of cells, tissues and even entire organisms. In this review, we discuss recent developments and applications of paFPs in super-resolution localization imaging.     

Challenges of manufacturing mesenchymal stromal cell–derived extracellular vesicles in regenerative medicine

The field of regenerative medicine has expanded greatly in the past decade, with more than 1000 current clinical trials involving mesenchymal stromal cell (MSC) treatment. Multiple recent publications have demonstrated that the beneficial effects from MSCs are not simply due to engraftment into the target organ as classically thought but rather are largely attributable to the release of paracrine factors including cytokines, growth factors and extracellular vesicles (EVs). These EVs contain miRNAs, free fatty acids and proteins that promote regeneration, proliferation and cell function and improve inflammation. Although EVs have shown promising results in animal studies, there are many obstacles to the manufacturing of EVs for clinical applications. This review discusses challenges associated with the manufacturing of clinical-grade EVs in regard to identity, purity, reproducibility, sterility, storage, potency and safety. We discuss currently employed methods and approaches for developing clinical Good Manufacturing Practices (GMP)-grade EVs and the limitations for each. We further discuss the best approaches to overcome the current hurdles in developing clinical GMP-grade EVs.     

The effects of simulated microgravity on avian embryonic development.

Based on the few reports available, microgravity (MG) can have adverse effects on the early development of vascularised extra-embryonic membranes in avian eggs. Whether gravity or oxygen availability is the stimulus for development of the blood vessels in the chorioallantoic membranes (CAM) remains unclear. Under gravity the blastoderm forms on top of the yolk sac, closest to the oxygen rich region beneath the shell membranes, and from there the CAM buds from an abdominal extension subsequently to form a close contact with shell membranes. Then as the embryo develops it spreads beneath the eggshell surface to maximise the surface area of the CAM vascular bed available for O2 uptake. To investigate how simulated MG influences development of the CAM and embryo we conducted experiments on chicken embryos during incubation in a 3D-clinostat (control or continuous MG treatment at 5 rpm). Further, to determine if CAM angiogenesis is directed towards regions of high O2 tension or gravity we investigated the effects of wax treatment (50% shell surface area) on development in MG. We found that clinostat MG caused embryonic failure between day 0-5 by preventing normal development of CAM-shell membrane complex. Thereafter acute MG promoted increases in CAM mass, but did not affect embryo mass. Preliminary findings suggest that combined acute MG and wax treatment did not significantly affect embryonic growth in either MG or control groups, but retarded CAM growth in control embryos only. Finally, we will present evidence to show that acute and prolonged exposure to MG does not prevent normal growth and hatching, but might have more subtle effects on hatchling physiology, including reduced heart mass.     

APPLICATION OF FUZZY SETS AND NEURAL NETWORKS IN SENSORY ANALYSIS

This paper describes an initial but fundamental attempt to lay some groundwork for a fuzzy-set-based paradigm for sensory analysis and to demonstrate how fuzzy set and neural network techniques may lead to a natural way for sensory data interpretation. Sensory scales are described as fuzzy sets, sensory attributes as fuzzy variables, and sensory responses as sample membership grades. Multi-judge responses are formulated as a fuzzy membership vector or fuzzy histogram of response, which gives an overall panel response free of the unverifiable assumptions implied in conventional approaches. Neural networks are used to provide an effective tool for modeling and analysis of sensory responses in their naturally fuzzy and complex forms. A maximum method of defuzzification is proposed to give a crisp grade of the majority opinion. Two applications in meat quality evaluation are used to demonstrate the use of the paradigm and procedure. It is hoped that this work will bring up some new ideas and generate interest in research on application of fuzzy sets and neural networks in sensory analysis.     

Data handling for interactive metabolomics: tools for studying the dynamics of metabolome-macromolecule interactions

All published metabolomics studies investigate changes in either absolute or relative quantities of metabolites. However, blood plasma, one of the most commonly studied biofluids for metabolomics applications, is a complex, heterogeneous mixture of lipoproteins, proteins, small organic molecules and ions which together undergo a variety of possible molecular interactions including metal complexation, chemical exchange processes, micellular compartmentation of metabolites, enzyme-mediated biotransformations and small-molecule-macromolecule binding. In particular, many low molecular weight (MW) compounds (including drugs) can exist both ‘free’ in solution and bound to proteins or within organised aggregates of macromolecules. To study the effects of e.g. disease on these interactions we suggest that new approaches are needed. We have developed a technique termed ‘interactive metabolomics’ or i-metabolomics. i-metabolomics can be defined as: “The study of interactions between low MW biochemicals and macromolecules in heterogeneous biosamples such as blood plasma, without pre-selection of the components of interest”. Standard 1D NMR experiments commonly used in metabolomics allow metabolite concentration differences between samples to be investigated because the intensity of each peak depends on the concentration of the compound in question. On the other hand, the instrument can be set-up to measure molecular interactions by monitoring the diffusion coefficients of molecules. According to the Stokes–Einstein equation, the diffusion coefficient of a molecule is inversely proportional to its effective size, as represented by the hydrodynamic radius. Therefore, when low MW compounds are non-covalently bound to proteins, the observed diffusion coefficient for the compound will be intermediate between those of its free and bound forms. By measuring diffusion by NMR, the degree of protein binding can be estimated for either low MW endogenous biochemicals or xenobiotics. This type of experiment is referred to as either Diffusion-Ordered Spectroscopy (DOSY) or Diffusion-Edited Spectroscopy, depending on the type of post-acquisition data processing applied to the spectra. Results presented in this paper demonstrate approaches for the non-selective modelling of metabolite-macromolecule interactions (i-metabolomics), whilst additionally highlighting some of the all too frequently ignored issues associated with interpretation of data derived from profiling of blood plasma.     

Model-based uncertainty in species range prediction

Aim Many attempts to predict the potential range of species rely on environmental niche (or ‘bioclimate envelope’) modelling, yet the effects of using different niche‐based methodologies require further investigation. Here we investigate the impact that the choice of model can have on predictions, identify key reasons why model output may differ and discuss the implications that model uncertainty has for policy‐guiding applications. Location The Western Cape of South Africa. Methods We applied nine of the most widely used modelling techniques to model potential distributions under current and predicted future climate for four species (including two subspecies) of Proteaceae. Each model was built using an identical set of five input variables and distribution data for 3996 sampled sites. We compare model predictions by testing agreement between observed and simulated distributions for the present day (using the area under the receiver operating characteristic curve (AUC) and kappa statistics) and by assessing consistency in predictions of range size changes under future climate (using cluster analysis). Results Our analyses show significant differences between predictions from different models, with predicted changes in range size by 2030 differing in both magnitude and direction (e.g. from 92% loss to 322% gain). We explain differences with reference to two characteristics of the modelling techniques: data input requirements (presence/absence vs. presence‐only approaches) and assumptions made by each algorithm when extrapolating beyond the range of data used to build the model. The effects of these factors should be carefully considered when using this modelling approach to predict species ranges. Main conclusions We highlight an important source of uncertainty in assessments of the impacts of climate change on biodiversity and emphasize that model predictions should be interpreted in policy‐guiding applications along with a full appreciation of uncertainty.