Search

[Display omitted] • Phylogenetic model selection can be performed using neural networks. • A resnet-18 neural network can be used to determine the model of sequence evolution. • A Bi-LSTM can be used to determine presence of rate heterogeneity and estimate the alpha parameter of the Γ --distribution. • Trees reconstructed using the models resulting from neural network estimates are closer to the ground truth than maximum likelihood trees. • Neural network computation times are constant, thus yielding strongly decreased runtimes for alignments with long sequences and many taxa. Selecting the best model of sequence evolution for a multiple-sequence-alignment (MSA) constitutes the first step of phylogenetic tree reconstruction. Common approaches for inferring nucleotide models typically apply maximum likelihood (ML) methods, with discrimination between models determined by one of several information criteria. This requires tree reconstruction and optimisation which can be computationally expensive. We demonstrate that neural networks can be used to perform model selection, without the need to reconstruct trees, optimise parameters, or calculate likelihoods. We introduce ModelRevelator, a model selection tool underpinned by two deep neural networks. The first neural network, NNmodelfind, recommends one of six commonly used models of sequence evolution, ranging in complexity from Jukes and Cantor to General Time Reversible. The second, NNalphafind, recommends whether or not a Γ -distributed rate heterogeneous model should be incorporated, and if so, provides an estimate of the shape parameter, ɑ. Users can simply input an MSA into ModelRevelator, and swiftly receive output recommending the evolutionary model, inclusive of the presence or absence of rate heterogeneity, and an estimate of ɑ. We show that ModelRevelator performs comparably with likelihood-based methods and the recently published machine learning method ModelTeller over a wide range of parameter settings, with significant potential savings in computational effort. Further, we show that this performance is not restricted to the alignments on which the networks were trained, but is maintained even on unseen empirical data. We expect that ModelRevelator will provide a valuable alternative for phylogeneticists, especially where traditional methods of model selection are computationally prohibitive. [ABSTRACT FROM AUTHOR]

The interaction of a wing/blade profile with turbulent inflow is one of the main sources of sound generation in turbomachines. There are already several noise prediction methodologies available which either appear not to account for the influence of geometrical and flow parameters on noise generation or have been able to account for a remarkably limited extent due to the requirement of high-performance computing for flow calculations. This leads to the motivation of the current paper, which presents a low-cost and easy-to-use noise prediction methodology based on the statistical modelling of the inflow turbulence and Lookup Table (LUT) approach for aeroacoustic design and optimisation. The development of the statistical method is divided into three parts; namely – i) calculating the background flow, ii) modelling of statistically optimised inflow disturbance, iii) computing the far-field sound pressure for individual vortex passages and superpose them linearly – this step involves repeated computation of identical vortex passages and can be therefore easily sped up using a database approach. In the framework of this work, a new approach to model the inflow turbulence using vortex particles characterised by shape functions, based on waveforms, is presented. The idea is to not conduct a time-dependent unsteady calculation of the flow field in real-time, instead to consider the mean flow around the profile in the computational domain, in which the vortex particles are convected to realise a statistical turbulent signal. The convection of these vortex particles, also, does not take place in the real-time calculation, instead, vortices of every possible size and strength are convected in a similar domain with a specific airfoil, and the acoustic radiation due to their interaction with the airfoil are computed and stored in a database. The far-field noise is predicted using Curle's formulation. The generated database is accessed using the LUT approach to rapidly extract the acoustic signals. Through this approach, the influence of geometrical as well as flow parameters on the noise generated by airfoils can be quantified without requiring to conduct a numerical simulation every time for a new set of geometrical and flow variables. In the article, the application of the method for different blade profiles is shown, and the results obtained are compared with the standard literature. [ABSTRACT FROM AUTHOR]

Abstract: Regeneration is widespread throughout the animal kingdom, but our molecular understanding of this process in adult animals remains poorly understood. Wnt/β-catenin signaling plays crucial roles throughout animal life from early development to adulthood. In intact and regenerating planarians, the regulation of Wnt/β-catenin signaling functions to maintain and specify anterior/posterior (A/P) identity. Here, we explore the expression kinetics and RNAi phenotypes for secreted members of the Wnt signaling pathway in the planarian Schmidtea mediterranea. Smed-wnt and sFRP expression during regeneration is surprisingly dynamic and reveals fundamental aspects of planarian biology that have been previously unappreciated. We show that after amputation, a wounding response precedes rapid re-organization of the A/P axis. Furthermore, cells throughout the body plan can mount this response and reassess their new A/P location in the complete absence of stem cells. While initial stages of the amputation response are stem cell independent, tissue remodeling and the integration of a new A/P address with anatomy are stem cell dependent. We also show that WNT5 functions in a reciprocal manner with SLIT to pattern the planarian mediolateral axis, while WNT11-2 patterns the posterior midline. Moreover, we perform an extensive phylogenetic analysis on the Smed-wnt genes using a method that combines and integrates both sequence and structural alignments, enabling us to place all nine genes into Wnt subfamilies for the first time. [ABSTRACT FROM AUTHOR]

Abstract: Wnt genes and β-Catenin signaling are involved in axial patterning processes in vertebrate embryogenesis in setting up the Spemann–Mangold organizer in amphibian embryos. An organizer with a similar function is present in the hypostome of an adult Hydra polyp. Previously, a Hydra ortholog of Wnt3 (HyWnt3), which is expressed in the hypostome, has been described. Here, ten additional Hydra Wnt genes have been identified. Of these, six (HyWnt1, -7, -9/10a, -9/10c, -11, and -16) are expressed in the adult hypostome. And, as is HyWnt3, these six Wnt genes are also expressed when a new head organizer is formed during head regeneration and bud formation. The kinetics of Wnt gene expressions during head regeneration suggests that a cascade of consecutive Wnt activation accompanies regeneration, and HyWnt3 begins this cascade. Recombinant HyWnt3 protein induced body column tissue to undergo head formation. It also increased the head formation capacity in the head regeneration-deficient mutant strain reg-16 to that of wild-type strains. In addition our data reveal striking similarities in the molecular basis of the organizer in Hydra and axis polarization in chordates (e.g. Spemann''s organizer) as well as it''s role in regeneration suggesting a conserved function of Wnt signaling in setting up this ancient metazoan signaling center. [Copyright &y& Elsevier]