Thinking of Fish Population Discrimination: Population Average Phenotype vs. Population Phenotypes

The genetic polymorphism and phenotypic variation are key in ecology and evolution. The morphological variability of the contour of fish otoliths has been extensively used for the delimitation of stocks. These studies are conventionally based on average phenotype using elliptic Fourier analysis and lineal discriminant analysis as classifier. Considering new analytical options, such as the wavelet transform and non-parametric algorithms, we here analyzed the otolith shape of Trachurus picturatus (blue jack mackerel) from mainland Portugal, Madeira, and the Canaries. We explore the phenotypic variation throughout a latitudinal gradient, establish a hypothesis to explain this variability based on the reaction norms, and determine how the use of average phenotype and/or morphotypes influences in the delimitation of stocks. Four morphotypes were identified in all regions, with an increase of phenotypes in warmer waters. The findings demonstrated that stocks were clearly separated with classification rates over 90%. The use of morphotypes, revealed seasonal variations in their frequencies and per region. The presence of shared phenotypes in different proportions among fishing grounds may open new management approaches in migratory species. These results show the importance of the phenotypic diversity in fisheries management.

521Performance of doubly-robust, machine learning effect estimators in realistic epidemiologic data settings and practical recommendations

Background Flexible, data-adaptive algorithms (machine learning; ML) for nuisance parameter estimation in epidemiologic causal inference have promising asymptotic properties for complex, high-dimensional data. However, recently proposed applications (e.g. targeted maximum likelihood estimation; TMLE) may produce biases parameter and standard error estimates in common real-world cohort settings. The relative performance of these novel estimators over simpler approaches in such settings is unclear. Methods We apply double-crossfit TMLE, augmented inverse probability weighting (AIPW), and standard IPW to simple simulations (5 covariates) and “real-world” data using covariate-structure-preserving (“plasmode”) simulations of 1,178 subjects and 331 covariates from a longitudinal birth cohort. We evaluate various data generating and estimation scenarios including: under- and over- (e.g. excess orthogonal covariates) identification, poor data support, near-instruments, and mis-specified biological interactions. We also track representative computation times. Results We replicate optimal performance of cross-fit, doubly robust estimators in simple data generating processes. However, in nearly every real world-based scenario, estimators fit with parametric learners outperform those that include non-parametric learners in terms of mean bias and confidence interval coverage. Even when correctly specified, estimators fit with non-parametric algorithms (xgboost, random forest) performed poorly (e.g. 24% bias, 57% coverage vs. 10% bias, 79% coverage for parametric fit), at times underperforming simple IPW. Conclusions In typical epidemiologic data sets, double-crossfit estimators fit with simple smooth, parametric learners may be the optimal solution, taking 2-5 times less computation time than flexible non-parametric models, while having equal or better performance. No approaches are optimal, and estimators should be compared on simulations close to the source data. Key messages In epidemiologic studies, use of flexible non-parametric algorithms for effect estimation should be strongly justified (i.e. high-dimensional covariates) and performed with care. Parametric learners may be a safer option with few drawbacks.

Advanced sun-sensor processing and design for super-resolution performance

The performance of conventional and parametric super-resolution algorithms for estimating sun position in a spacecraft sun-sensor was analyzed. Widely employed in other applications, parametric algorithms were examined to evaluate increase in system performance without affecting the cost of the sensor system. Using a simplified model of detector illumination simulations provided quantitative comparisons of algorithm performance. Simple sensor re-design was examined by using genetic algorithms as a heuristic to optimize the illumination pattern for a single axis digital sun-sensor. Findings show that, multiple narrow peak patterns provide subpixel accuracy in resolving the sun-angle. The optimal illumination pattern can be implemented by fabricating a replacement aperture mask for the sensor and this change can be made at a minimal cost. The super-resolution algorithms were tested with a component noise model and image degradation due to Earth albedo effects were examined. Parametric algorithms display very good performance throughout the test regime. The improvements are substantial enough to validate this approach worthy of future study.

Advanced sun-sensor processing and design for super-resolution performance

The performance of conventional and parametric super-resolution algorithms for estimating sun position in a spacecraft sun-sensor was analyzed. Widely employed in other applications, parametric algorithms were examined to evaluate increase in system performance without affecting the cost of the sensor system. Using a simplified model of detector illumination simulations provided quantitative comparisons of algorithm performance. Simple sensor re-design was examined by using genetic algorithms as a heuristic to optimize the illumination pattern for a single axis digital sun-sensor. Findings show that, multiple narrow peak patterns provide subpixel accuracy in resolving the sun-angle. The optimal illumination pattern can be implemented by fabricating a replacement aperture mask for the sensor and this change can be made at a minimal cost. The super-resolution algorithms were tested with a component noise model and image degradation due to Earth albedo effects were examined. Parametric algorithms display very good performance throughout the test regime. The improvements are substantial enough to validate this approach worthy of future study.

Unsupervised multi-instance learning for protein structure determination

Many regions of the protein universe remain inaccessible by wet-laboratory or computational structure determination methods. A significant challenge in elucidating these dark regions in silico relates to the ability to discriminate relevant structure(s) among many structures/decoys computed for a protein of interest, a problem known as decoy selection. Clustering decoys based on geometric similarity remains popular. However, it is unclear how exactly to exploit the groups of decoys revealed via clustering to select individual structures for prediction. In this paper, we provide an intuitive formulation of the decoy selection problem as an instance of unsupervised multi-instance learning. We address the problem in three stages, first organizing given decoys of a protein molecule into bags, then identifying relevant bags, and finally drawing individual instances from these bags to offer as prediction. We propose both non-parametric and parametric algorithms for drawing individual instances. Our evaluation utilizes two datasets, one benchmark dataset of ensembles of decoys for a varied list of protein molecules, and a dataset of decoy ensembles for targets drawn from recent CASP competitions. A comparative analysis with state-of-the-art methods reveals that the proposed approach outperforms existing methods, thus warranting further investigation of multi-instance learning to advance our treatment of decoy selection.

Symmetric Free Form Building Structures Arranged Regularly on Smooth Surfaces with Polyhedral Nets

The article is an original insight into interdisciplinary challenges of shaping innovative unconventional complex free form buildings roofed with multi-segment shell structures arranged with using novel parametric regular networks. The roof structures are made up of nominally plane thin-walled folded steel sheets transformed elastically and rationally into spatial shapes. A method is presented for creating such symmetric structures based on the regular spatial polyhedral networks created as a result of a composition of many complete reference tetrahedrons by their common flat sides and straight side edges arranged regularly and symmetrically in the three-dimensional Euclidean space. The use of the regularity and symmetry in the process of shaping different forms of (a) single tetrahedral meshes and whole consistent polyhedral structures, (b) individual plane walls and complex elevations, (c) single transformed folds, entire corrugated shell roofs, and their structures allow a creative search for attractive rational parametric solutions using a few author’s parametric algorithms and their implementation as built-in commands of the AutoCAD visual editor or applications of the Rhino/Grasshopper program.