Search for: All records

Just as a phylogeny encodes the evolutionary relationships among a group of organisms, a cophylogeny represents the coevolutionary relationships among symbiotic partners. Both are primarily reconstructed using computational analysis of biomolecular sequence data. The most widely used cophylogenetic reconstruction methods utilize an important simplifying assumption: species phylogenies for each set of coevolved taxa are required as input and assumed to be correct. Many studies have shown that this assumption is rarely – if ever – satisfied, and the consequences for cophylogenetic studies are poorly understood. To address this gap, we conduct a comprehensive performance study that quantifies the relationship between species tree estimation error and downstream cophylogenetic estimation accuracy. We study the performance of state-of-the-art methods for cophylogenetic reconstruction using in silico model-based simulations. Our investigation also assessed cophylogenetic reproducibility using genomic sequence data from two important models of symbiosis: soil-associated fungi and their endosymbiotic bacteria, and bobtail squid and their bioluminescent bacterial symbionts. Our findings conclusively demonstrate the major impact that upstream phylogenetic estimation error has on downstream cophylogenetic reconstruction. Relative to other experimental factors such as cophylogenetic estimation method choice and coevolutionary event costs, phylogenetic estimation error ranked highest in importance based on a random forest-based variable importance assessment. We conclude with practical guidance and future research directions. Among the many considerations needed for accurate cophylogenetic reconstruction – choice of computational method, method settings, sampling design, and others – just as much attention must be paid to careful species phylogeny estimation using modern best practices. 

Advances in algorithms and low-power computing hardware imply that machine learning is of potential use in off-grid medical data classification and diagnosis applications such as electrocardiogram interpretation. However, although support vector machine algorithms for electrocardiogram classification show high classification accuracy, hardware implementations for edge applications are impractical due to the complexity and substantial power consumption needed for kernel optimization when using conventional complementary metal–oxide–semiconductor circuits. Here we report reconfigurable mixed-kernel transistors based on dual-gated van der Waals heterojunctions that can generate fully tunable individual and mixed Gaussian and sigmoid functions for analogue support vector machine kernel applications. We show that the heterojunction-generated kernels can be used for arrhythmia detection from electrocardiogram signals with high classification accuracy compared with standard radial basis function kernels. The reconfigurable nature of mixed-kernel heterojunction transistors also allows for personalized detection using Bayesian optimization. A single mixed-kernel heterojunction device can generate the equivalent transfer function of a complementary metal–oxide–semiconductor circuit comprising dozens of transistors and thus provides a low-power approach for support vector machine classification applications.