Deep forest ensemble learning for classification of alignments of non-coding RNA sequences based on multi-view structure representations

Non-coding RNAs (ncRNAs) play crucial roles in multiple biological processes. However, only a few ncRNAs’ functions have been well studied. Given the significance of ncRNAs classification for understanding ncRNAs’ functions, more and more computational methods have been introduced to improve the classification automatically and accurately. In this paper, based on a convolutional neural network and a deep forest algorithm, multi-grained cascade forest (GcForest), we propose a novel deep fusion learning framework, GcForest fusion method (GCFM), to classify alignments of ncRNA sequences for accurate clustering of ncRNAs. GCFM integrates a multi-view structure feature representation including sequence-structure alignment encoding, structure image representation and shape alignment encoding of structural subunits, enabling us to capture the potential specificity between ncRNAs. For the classification of pairwise alignment of two ncRNA sequences, the F-value of GCFM improves 6% than an existing alignment-based method. Furthermore, the clustering of ncRNA families is carried out based on the classification matrix generated from GCFM. Results suggest better performance (with 20% accuracy improved) than existing ncRNA clustering methods (RNAclust, Ensembleclust and CNNclust). Additionally, we apply GCFM to construct a phylogenetic tree of ncRNA and predict the probability of interactions between RNAs. Most ncRNAs are located correctly in the phylogenetic tree, and the prediction accuracy of RNA interaction is 90.63%. A web server (http://bmbl.sdstate.edu/gcfm/) is developed to maximize its availability, and the source code and related data are available at the same URL.

Computational geometric tools for quantitative comparison of locomotory behavior

A fundamental challenge for behavioral neuroscientists is to accurately quantify (dis)similarities in animal behavior without excluding inherent variability present between individuals. We explored two new applications of curve and shape alignment techniques to address this issue. As a proof-of-concept we applied these methods to compare normal or alarmed behavior in pairs of medaka (Oryzias latipes). The curve alignment method we call Behavioral Distortion Distance (BDD) revealed that alarmed fish display less predictable swimming over time, even if individuals incorporate the same action patterns like immobility, sudden changes in swimming trajectory, or changing their position in the water column. The Conformal Spatiotemporal Distance (CSD) technique on the other hand revealed that, in spite of the unpredictability, alarmed individuals exhibit lower variability in overall swim patterns, possibly accounting for the widely held notion of “stereotypy” in alarm responses. More generally, we propose that these new applications of established computational geometric techniques are useful in combination to represent, compare, and quantify complex behaviors consisting of common action patterns that differ in duration, sequence, or frequency.

Geometric Deep Learning for Shape Correspondence in Mass Customization

Many industries, such as human-centric product manufacturers, are calling for mass customization with personalized products. One key enabler of mass customization is 3D printing, which makes the flexible design and manufacturing possible. However, personalized designs bring obstacles for the shape matching and analysis, owing to the high complexity and large shape variations. Traditional shape matching methods are limited to shape alignment, which cannot determine the intrinsic in-variance of mass customized models. To extract the deformations widely seen in mass customization paradigm and address the issues of alignment methods in shape matching, we redefine the geometry matching problem as a correspondence problem, and solve for the correspondence of all vertices on a queried shape to a reference shape. A state-of-the-art geometric deep learning method is used to learn the correspondence of a set of collected models. Through learning the intrinsic deformations of the products, the underlying variations of the shapes are extracted. We demonstrate the application of the proposed approach in orthodontics industry, and the experimental results show the effectiveness of the proposed method and the defined problem is favorably suitable for shape analysis in mass customization.

The intricate link between galaxy dynamics and intrinsic shape (or why so-called prolate rotation is a misnomer)

Many recent integral field spectroscopy (IFS) survey teams have used stellar kinematic maps combined with imaging to statistically infer the underlying distributions of galaxy intrinsic shapes. With now several IFS samples at our disposal, the method, which was originally proposed by M. Franx and collaborators in 1991, is gaining in popularity, having been so far applied to ATLAS3D, SAMI, MANGA and MASSIVE. We present results showing that a commonly assumed relationship between dynamical and intrinsic shape alignment does not hold in Illustris, affecting our ability to recover accurate intrinsic shape distributions. A further implication is that so-called “prolate rotation”, where the bulk of stars in prolate galaxies are thought to rotate around the projected major axis, is a misnomer.

Computational Geometric Tools for Modeling Inherent Variability in Animal Behavior

ABSTRACTA fundamental challenge for behavioral neuroscientists is to represent inherent variability among animals accurately without compromising the ability to quantify differences between conditions. We developed two new methods that apply curve and shape alignment techniques to address this issue. As a proof-of-concept we applied these methods to compare normal or alarmed behavior in pairs of medaka (Oryzias latipes). The curve alignment method we call Behavioral Distortion Distance (BDD) revealed that alarmed fish display less predictable swimming over time, even if individuals incorporate the same action patterns like immobility, sudden changes in swimming trajectory, or changing their position in the water column. The Conformal Spatiotemporal Distance (CSD) technique on the other hand revealed that, in spite of the unpredictability, alarmed individuals share an overall swim pattern, possibly accounting for the widely held notion of “stereotypy” in alarm responses. More generally, we propose that these new applications of known computational geometric techniques are useful in combination to represent, compare, and quantify complex behaviors consisting of common action patterns that differ in duration, sequence, or frequency.