DD-Classifier: Nonparametric Classification Procedure Based onDD-Plot

A review of model evaluation metrics for machine learning in genetics and genomics

Machine learning (ML) has shown great promise in genetics and genomics where large and complex datasets have the potential to provide insight into many aspects of disease risk, pathogenesis of genetic disorders, and prediction of health and wellbeing. However, with this possibility there is a responsibility to exercise caution against biases and inflation of results that can have harmful unintended impacts. Therefore, researchers must understand the metrics used to evaluate ML models which can influence the critical interpretation of results. In this review we provide an overview of ML metrics for clustering, classification, and regression and highlight the advantages and disadvantages of each. We also detail common pitfalls that occur during model evaluation. Finally, we provide examples of how researchers can assess and utilise the results of ML models, specifically from a genomics perspective.

Uncertainty analysis of contagion processes based on a functional approach

The spread of a disease, product or idea in a population is often hard to predict. We tend to observe one or few realizations of the contagion process and therefore limited information can be obtained for anticipating future similar events. The stochastic nature of contagion generates unpredictable outcomes throughout the whole course of the dynamics. This might lead to important inaccuracies in the predictions and to the over or under-reaction of policymakers, who tend to anticipate the average behavior. Through an extensive simulation study, we analyze properties of the contagion process, focusing on its unpredictability or uncertainty, and exploiting the functional nature of the data. In particular, we define a novel non-parametric measure of variance based on weighted depth-based central regions. We apply this methodology to the susceptible-infected-susceptible epidemiological model and small-world networks. We find that maximum uncertainty is attained at the epidemic threshold. The density of the network and the contagiousness of the process have a strong and complementary effect on the uncertainty of contagion, whereas only a mild effect of the network's randomness structure is observed.

Depth-based reconstruction method for incomplete functional data

The problem of estimating missing fragments of curves from a functional sample has been widely considered in the literature. However, most reconstruction methods rely on estimating the covariance matrix or the components of its eigendecomposition, which may be difficult. In particular, the estimation accuracy might be affected by the complexity of the covariance function, the noise of the discrete observations, and the poor availability of complete discrete functional data. We introduce a non-parametric alternative based on depth measures for partially observed functional data. Our simulations point out that the benchmark methods perform better when the data come from one population, curves are smooth, and there is a large proportion of complete data. However, our approach is superior when considering more complex covariance structures, non-smooth curves, and when the proportion of complete functions is scarce. Moreover, even in the most severe case of having all the functions incomplete, our method provides good estimates; meanwhile, the competitors are unable. The methodology is illustrated with two real data sets: the Spanish daily temperatures observed in different weather stations and the age-specific mortality by prefectures in Japan. They highlight the interpretability potential of the depth-based method.

Tukey's Depth for Object Data

We develop a novel exploratory tool for non-Euclidean object data based on data depth, extending celebrated Tukey's depth for Euclidean data. The proposed metric halfspace depth, applicable to data objects in a general metric space, assigns to data points depth values that characterize the centrality of these points with respect to the distribution and provides an interpretable center-outward ranking. Desirable theoretical properties that generalize standard depth properties postulated for Euclidean data are established for the metric halfspace depth. The depth median, defined as the deepest point, is shown to have high robustness as a location descriptor both in theory and in simulation. We propose an efficient algorithm to approximate the metric halfspace depth and illustrate its ability to adapt to the intrinsic data geometry. The metric halfspace depth was applied to an Alzheimer's disease study, revealing group differences in the brain connectivity, modeled as covariance matrices, for subjects in different stages of dementia. Based on phylogenetic trees of 7 pathogenic parasites, our proposed metric halfspace depth was also used to construct a meaningful consensus estimate of the evolutionary history and to identify potential outlier trees.

NITPicker: selecting time points for follow-up experiments

Background

The design of an experiment influences both what a researcher can measure, as well as how much confidence can be placed in the results. As such, it is vitally important that experimental design decisions do not systematically bias research outcomes. At the same time, making optimal design decisions can produce results leading to statistically stronger conclusions. Deciding where and when to sample are among the most critical aspects of many experimental designs; for example, we might have to choose the time points at which to measure some quantity in a time series experiment. Choosing times which are too far apart could result in missing short bursts of activity. On the other hand, there may be time points which provide very little information regarding the overall behaviour of the quantity in question.

Results

In this study, we develop a tool called NITPicker (Next Iteration Time-point Picker) for selecting optimal time points (or spatial points along a single axis), that eliminates some of the biases caused by human decision-making, while maximising information about the shape of the underlying curves. NITPicker uses ideas from the field of functional data analysis. NITPicker is available on the Comprehensive R Archive Network (CRAN) and code for drawing figures is available on Github (https://github.com/ezer/NITPicker).

Conclusions

NITPicker performs well on diverse real-world datasets that would be relevant for varied biological applications, including designing follow-up experiments for longitudinal gene expression data, weather pattern changes over time, and growth curves.

Electronic supplementary material

The online version of this article (10.1186/s12859-019-2717-5) contains supplementary material, which is available to authorized users.

Air quality across a European hotspot: Spatial gradients, seasonality, diurnal cycles and trends in the Veneto region, NE Italy

The Veneto region (NE Italy) lies in the eastern part of the Po Valley, a European hotspot for air pollution. Data for key air pollutants (CO, NO, NO₂, O₃, SO₂, PM₁₀ and PM_2.5) measured over 7years (2008/2014) across 43 sites in Veneto were processed to characterise their spatial and temporal patterns and assess the air quality. Nitrogen oxides, PM and ozone are critical pollutants frequently breaching the EC limit and target values. Intersite analysis demonstrates a widespread pollution across the region and shows that primary pollutants (nitrogen oxides, CO, PM) are significantly higher in cities and over the flat lands due to higher anthropogenic pressures. The spatial variation of air pollutants at rural sites was then mapped to depict the gradient of background pollution: nitrogen oxides are higher in the plain area due to the presence of strong diffuse anthropogenic sources, while ozone increases toward the mountains probably due to the higher levels of biogenic ozone-precursors and low NO emissions which are not sufficient to titrate out the photochemical O₃. Data-depth classification analysis revealed a poor categorization among urban, traffic and industrial sites: weather and urban planning factors may cause a general homogeneity of air pollution within cities driving this poor classification. Seasonal and diurnal cycles were investigated: the effect of primary sources in populated areas is evident throughout the region and drives similar patterns for most pollutants: road traffic appears the predominant potential source shaping the daily cycles. Trend analysis of experimental data reveals a general decrease of air pollution across the region, which agrees well with changes assessed by emission inventories. This study provides key information on air quality across NE Italy and highlights future research needs and possible developments of the regional monitoring network.