Automated insect identification systems have been explored for more than two decades but have only recently started to take advantage of powerful and versatile convolutional neural networks (CNNs). While typical CNN applications still require large training image datasets with hundreds of images per taxon, pretrained CNNs recently have been shown to be highly accurate, while being trained on much smaller datasets. We here evaluate the performance of CNN-based machine learning approaches in identifying three curated species-level dorsal habitus datasets for Miridae, the plant bugs. Miridae are of economic importance, but species-level identifications are challenging and typically rely on information other than dorsal habitus (e.g., host plants, locality, genitalic structures). Each dataset contained 2–6 species and 126–246 images in total, with a mean of only 32 images per species for the most difficult dataset. We find that closely related species of plant bugs can be identified with 80–90% accuracy based on their dorsal habitus alone. The pretrained CNN performed 10–20% better than a taxon expert who had access to the same dorsal habitus images. We find that feature extraction protocols (selection and combination of blocks of CNN layers) impact identification accuracy much more than the classifying mechanism (support vector machine and deep neural network classifiers). While our network has much lower accuracy on photographs of live insects (62%), overall results confirm that a pretrained CNN can be straightforwardly adapted to collection-based images for a new taxonomic group and successfully extract relevant features to classify insect species.
Despite many bioinformatic solutions for analyzing sequencing data, few options exist for targeted sequence retrieval from whole genomic sequencing (WGS) data with the ultimate goal of generating a phylogeny. Available tools especially struggle at deep phylogenetic levels and necessitate amino-acid space searches, which may increase rates of false positive results. Many tools are also difficult to install and may lack adequate user resources. Here, we describe a program that uses freely available similarity search tools to find homologs in assembled WGS data with unparalleled freedom to modify parameters. We evaluate its performance compared to other commonly used bioinformatics tools on two divergent insect species (>200 My) for which annotated genomes exist, and on one large set each of highly conserved and more variable loci. Our software is capable of retrieving orthologs from well-curated or unannotated, low or high depth shotgun, and target capture assemblies as well or better than other software as assessed by recovering the most genes with maximal coverage and with a low rate of false positives throughout all datasets. When assessing this combination of criteria, ALiBaSeq is frequently the best evaluated tool for gathering the most comprehensive and accurate phylogenetic alignments on all types of data tested. The software (implemented in Python), tutorials, and manual are freely available at https://github.com/AlexKnyshov/alibaseq.
specimen, for specimens at least as old as ~1965, and can be easily conducted in most laboratories.4. If widely applied, this technique will accelerate the accurate resolution of the Tree of Life especially on non-model organisms with limited existing genomic resources.
A cladistic analysis of the tribe Bryocorini based on 68 morphological characters is conducted. Bryocorini are supported as a monophyletic group with Eccritotarsini as their sister taxon. Based on the phylogenetic analysis, we redefine the tribe Bryocorini to contain the following seven genera: Bryocorella Carvalho, 1956, Bryocoris Fallén, 1829, Bryophilocapsus Yasunaga, 2000, Cobalorrhynchus Reuter, 1906 gen. dist., Diplazicoris gen. nov., Hekista Kirkaldy, 1902, and Monalocoris Dahlbom, 1851. The genus Bryocorella is transferred to Bryocorini from the tribe Eccritotarsini. The subgenus Cobalorrhynchus is treated as a separate genus. Diplazicoris is described as monotypic to accommodate Diplazicoris lombokianus sp. nov. An updated diagnosis of the tribe, a key to genera, and a diagnosis of each recognized genus are presented. Selected photomicrographs, scanning micrographs, and illustrations of the pretarsus, metepisternal scent efferent system, metafemoral trichobothria, and morphology of head, pronotum, and male and female genitalia are provided. Mapping of the host data on the revealed tree shows that Bryocorini represent one of the very few currently known examples of the adaptive radiation of a fairly large insect group on ferns.
Insect male genitalia show an evolutionarily variable morphology that has proven to be valuable for both, species identifications and phylogenetic analyses at higher taxonomic levels. Accurate usage of genitalic characters in taxonomic descriptions and phylogenetic analyses depends on consistency of terminology and validity of homology hypotheses. Both areas are underdeveloped in many insect groups. We here document the morphology and advance homology hypotheses of male genitalic features for the hemipteran infraorder Dipsocoromorpha, the minute litter bugs. Genitalic structures and the pregenital abdomen in Dipsocoromorpha are strikingly modified and diverse compared to other Heteroptera. In addition to variation in the shape of phallic structures (parameres and aedeagus), minute litter bug genitalia vary in the direction and degree of asymmetry and feature a plethora of processes derived from various abdominal segments with significant variation at low taxonomic levels. Here, male genitalic structures for an extensive taxonomic sample (32 genera and 71 specimens) are documented using scanning electron and confocal microscopy, and a universal terminology for genitalic structures across minute litter bugs is established that will facilitate species discovery and evolutionary research. We conclude by proposing primary homology hypotheses across the infraorder that now can be tested in a phylogenetic framework.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.