Accurately identifying phage-host relationships from their genome sequences is still challenging, especially for those phages and hosts with less homologous sequences. In this work, focusing on identifying the phage-host relationships at the species and genus level, we propose a contrastive learning based approach to learn whole-genome sequence embeddings that can take account of phage-host interactions (PHIs). Contrastive learning is used to make phages infecting the same hosts close to each other in the new representation space. Specifically, we rephrase whole-genome sequences with frequency chaos game representation (FCGR) and learn latent embeddings that 'encapsulate' phages and host relationships through contrastive learning. The contrastive learning method works well on the imbalanced dataset. Based on the learned embeddings, a proposed pipeline named CL4PHI can predict known hosts and unseen hosts in training. We compare our method with two recently proposed state-of-the-art learning-based methods on their benchmark datasets. The experiment results demonstrate that the proposed method using contrastive learning improves the prediction accuracy on known hosts and demonstrates a zero-shot prediction capability on unseen hosts. In terms of potential applications, the rapid pace of genome sequencing across different species has resulted in a vast amount of whole-genome sequencing data that require efficient computational methods for identifying phage-host interactions. The proposed approach is expected to address this need by efficiently processing whole-genome sequences of phages and prokaryotic hosts and capturing features related to phage-host relationships for genome sequence representation. This approach can be used to accelerate the discovery of phage-host interactions and aid in the development of phage-based therapies for infectious diseases.

SUMMARY: The Kyoto Encyclopedia of Genes and Genomes (KEGG) database serves as a valuable systems biology resource and is widely utilized in diverse research fields. However, existing software does not allow flexible visualization and network analyses of the vast and complex KEGG data. We developed ggkegg, an R package that integrates KEGG information with ggplot2 and ggraph. ggkegg enables enhanced visualization and network analyses of KEGG data. We demonstrate the utility of the package by providing examples of its application in single-cell, bulk transcriptome, and microbiome analyses. ggkegg may empower researchers to analyze complex biological networks and present their results effectively. AVAILABILITY AND IMPLEMENTATION: The package and user documentation are available at: https://github.com/noriakis/ggkegg.

MOTIVATION: Bacteriophages/phages are the viruses that infect and replicate within bacteria and archaea, and rich in human body. To investigate the relationship between phages and microbial communities, the identification of phages from metagenome sequences is the first step. Currently, there are two main methods for identifying phages: database-based (alignment-based) methods and alignment-free methods. Database-based methods typically use a large number of sequences as references; alignment-free methods usually learn the features of the sequences with machine learning and deep learning models. RESULTS: We propose INHERIT which uses a deep representation learning model to integrate both database-based and alignment-free methods, combining the strengths of both. Pre-training is used as an alternative way of acquiring knowledge representations from existing databases, while the BERT-style deep learning framework retains the advantage of alignment-free methods. We compare INHERIT with four existing methods on a third-party benchmark dataset. Our experiments show that INHERIT achieves a better performance with the F1-score of 0.9932. In addition, we find that pre-training two species separately helps the non-alignment deep learning model make more accurate predictions. AVAILABILITY AND IMPLEMENTATION: The codes of INHERIT are now available in: https://github.com/Celestial-Bai/INHERIT. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Pedestrians have difficulty noticing hybrid vehicles (HVs) and electrical vehicles (EVs) quietly approaching from behind. We propose a vehicle detection scheme using a smartphone carried by a pedestrian. A notification of a vehicle approaching can be delivered to wearable devices such as Google Glass. We exploit the high-frequency switching noise generated by the motor unit in HVs and EVs. Although people are less sensitive to these high-frequency ranges, these sounds are prominent even on a busy street, and it is possible for a smartphone to detect these signals. The ambient sound captured at 48 kHz is converted to a feature vector in the frequency domain. A J48 classifier implemented on a smartphone can determine whether an EV or HV is approaching. We have collected a large amount of vehicle data at various locations. The false-positive and false-negative rates of our detection scheme are 1.2% and 4.95%, respectively. The first alarm was detected as early as 11.6 s before the vehicle approached the observer. The scheme can also determine the vehicle speed and vehicle type.

Aiming at developing a methodology for constructing continuous-time chaotic dynamical systems as flexible pattern generators, this paper discusses a strategy for binding desired unstable periodic orbits into a chaotic attractor. The strategy is comprised of the following two stages: constructing an interim “chaos-generating template”, and deforming the template according to the specifically desired orbits. INTRODUCTION Chaos is a fascinating phenomenon arising from nonlinearities in dynamical systems, and investigations on its applications to intelligent and flexible systems have appeared in many fields. An important aspect of chaos is that chaotic attractors embed or “host” an infinite number of unstable periodic orbits (UPO’s) bifurcated from prechaotic states (Ott, 2002). Among them, some distinctive orbits can be used for characterization or control purposes. For example, a variety of chaos control methods (Ott et al., 1990; Pyragas, 1992; Zhang et al., 2009) can stabilize UPO’s embedded in chaotic attractors, enlarging the operation range and/or enhancing the functionality of the system. In more recent years, the synthesis of chaos from various approaches (Chen and Ueta, 2002; Zelinka et al., 2008; Munoz-Pacheco and Tlelo-Cuautle, 2009) has been an active direction of research along the line of exploiting chaos. Primary concerns of these efforts include statistical and topological characteristics (e.g., invariant measure, Lyapunov spectrum, novel scrolling behaviors) that would be important in designing chaos-based information processing and communication applications. While the present study share some common motivation with the above-mentioned studies, we have put more focus on the geometrical shape and dynamical properties of UPO’s themselves from the viewpoint of the adaptive generation of periodic behaviors. Here our intention lies in extending the functionality of (stable) periodic pattern generators based on function approximation of vector fields, e.g., polynomial approximation (Okada and Nakamura, 2002) and neural network learning (Kuroe and Miura, 2006). In this paper, we consider continuous-time chaotic attractors as a container of UPO’s (patterns) where they can be stabilized, entrained, or targeted by external inputs into the dynamical system. In what follows, we propose a design strategy for binding desired UPO’s into a chaotic attractor governed by a polynomial vector field. The strategy is comprised of the following two stages: constructing an interim “chaos-generating template”, and deforming the template according to the specifically desired orbits. POLYNOMIAL VECTOR FIELDS We consider polynomial dynamical systems of the form ẋ = f(x)(x ∈ R ) where the vector field f(x) is represented as f(x) = [f1(x) · · · fN (x)] T = Φ(a(p1p2···pN ))θ(x), (1) θ(x) = [x1 · · · x ` N x `−1 1 x2 · · · 1] T . (2) Here, ` denotes the maximum degree of the polynomials, and the matrix Φ is comprised of the coefficients a(p1p2···pN ) of the polynomials. For example, with 3rdorder polynomials, the first element of a two-dimensional vector field is represented as f1(x) = a1(30)x 3 1 + a1(21)x 2 1x2 + a1(12)x1x 2 2 + a1(03)x 3 2 + a1(20)x 2 1 + a1(11)x1x2 + a1(02)x 2 2 + a1(10)x1 + a1(01)x2 + a1(00).(3) In our design process of the dynamical system, the coefficients are obtained by least-square fitting. To this end, we set up target vectors f(ηi) (i = 1, 2, · · · , L) at design points ηi on and in the vicinity of the target orbits, and construct matrices F = [f(η1) f(η2) · · · f(ηL)], (4) Θ = [θ(η1) θ(η2) · · · θ(ηL)]. (5) With these matrices, the least-square solution for Φ is given by Φ(a(p1 p2 ··· pN )) = FΘ # (6) where Θ is the Moore-Penrose pseudo inverse of Θ. Proceedings 24th European Conference on Modelling and Simulation ©ECMS Andrzej Bargiela, Sayed Azam Ali David Crowley, Eugene J.H. Kerckhoffs (Editors) ISBN: 978-0-9564944-0-5 / ISBN: 978-0-9564944-1-2 (CD)