Deep learning techniques have recently found application in biodiversity research. Mayflies (Ephemeroptera), stoneflies (Plecoptera) and caddisflies (Trichoptera), often abbreviated as EPT, are ...frequently used for freshwater biomonitoring due to their large numbers and sensitivity to environmental changes. However, the morphological identification of EPT species is a challenging but fundamental task. Morphological identification of these freshwater insects is therefore not only extremely time-consuming and costly, but also often leads to misjudgments or generates datasets with low taxonomic resolution. Here, we investigated the application of deep learning to increase the efficiency and taxonomic resolution of biomonitoring programs. Our database contains 90 EPT taxa (genus or species level), with the number of images per category ranging from 21 to 300 (16,650 in total). Upon completion of training, a CNN (Convolutional Neural Network) model was created, capable of automatically classifying these taxa into their appropriate taxonomic categories with an accuracy of 98.7 %. Our model achieved a perfect classification rate of 100 % for 68 of the taxa in our dataset. We achieved noteworthy classification accuracy with morphologically closely related taxa within the training data (e.g., species of the genus Baetis, Hydropsyche, Perla). Gradient-weighted Class Activation Mapping (Grad-CAM) visualized the morphological features responsible for the classification of the treated species in the CNN models. Within Ephemeroptera, the head was the most important feature, while the thorax and abdomen were equally important for the classification of Plecoptera taxa. For the order Trichoptera, the head and thorax were almost equally important. Our database is recognized as the most extensive aquatic insect database, notably distinguished by its wealth of included categories (taxa). Our approach can help solve long-standing challenges in biodiversity research and address pressing issues in monitoring programs by saving time in sample identification.
Display omitted
•EPT taxa indicate ecosystem health, but identification challenges hinder their use.•We use deep learning to improve biomonitoring efficiency and taxonomic resolution.•Convolutional Neural Networks achieved 98.7 % accuracy in classifying 90 EPT taxa.•This approach may replace conventional morphological identification methods.
Macrozoobenthic samples of two North Patagonian river basins, Negro (Argentina) and Valdivia (Chile) were taken in the period 1977-1983 in order to compare habitats of aquatic insects and other ...invertebrates on both slopes of the Andes and at the same latitude. Ten genera of Plecoptera were recorded by the author as nymphs: Gripopterygidae (Antarctoperla, Limnoperla, Potamoperla, Pelurgoperla , and Araucanioperla ?), Diamphipnoidae (Diamphipnoa and Diamphipnopsis ), Austroperlidae (Klapopteryx ), Notonemouridae (Austronemoura ) and Perlidae (Kempnyella ). Most of these genera of Plecoptera were collected in the upper rhithral. Only Antarctoperla and Limnoperla were found in reservoirs, in one lake and in medium-size to wide rivers. Gripoperygids were more frequent on the Argentine side; the other families were more abundant on the opposite slope; this situation can probably be explained by the different ecological conditions brought about by cattle raising and dam construction.
Fifteen stonefly species were found in the streams of the Swietokrzyski National Park (Lysogory chain of the Swietokrzyskie Mts); among them Protonemura intricata, Pr. praecox and Leuctra hippopus ...were found for the first time in this area. Clear differences between the stonefly fauna of the northern and southern slopes were observed.
The current knowledge of the Patagonian Plecoptera is due to the work by J. lilies during last Century's 50's and 60's, as well as to intermitent previous and posterior contributions. This knowledge ...is reasonable for the northern and central Patagonian Andes, and poor for the southernmost mountain areas (Santa Cruz and Tierra del Fuego provinces) and the steppe. New records are provided from Argentina. A list of the 83 species and 37 genera of Patagonian Plecoptera is given, out of which 47 species belonging to 28 genera are known to occur in Argentina.
The egg structures of five antarctoperlarian species – Stenoperla prasina of Eustheniidae; Austroperla cyrene of Austroperlidae; and Zelandobius truncus, Megaleptoperla grandis, and Acroperla ...trivacuata of Gripopterygidae, were examined in detail, and the groundplan of the egg structure was considered within the representative lineages of Antarctoperlaria and Plecoptera. The flattened egg shape and the circular arrangement of micropyles along the equator are regarded as potential autapomorphies for not only Eustheniidae but also for Eusthenioidea. Austroperlidae has eggs with thin, less-sclerotized chorion, a gelatinous layer on the surface, and micropyles roughly and randomly arranged along the equator. A significant ultrastructural difference between the attachment disc in Gripopterygidae and the anchor plate of arctoperlarian Systellognatha suggests that these structures were independently derived. The thin less-sclerotized chorion represents a groundplan feature in Plecoptera, along with micropyles arranged in a circle, including those circularly arranged along the equator of the egg. On the other hand, in contrast to previous understanding, the sclerotized hard chorion is regarded as a derived feature, having been independently acquired in each of Eustheniidae and Gripopterygidae of Antarctoperlaria and Systellognatha of Arctoperlaria.
•Egg structures of five Antarctoperlaria species were examined in detail.•Groundplan of egg structure in Antarctoperlaria and Plecoptera was discussed.•Antarctoperlarian attachment disc and arctoperlarian anchor plate were derived in parallel.•Not sclerotized chorion but thin less-sclerotized one represents a groundplan feature in Plecoptera.•Circular arrangement of micropyles is a groundplan feature of Plecoptera.
Display omitted
•Denser mitogenome sampling solved higher-level relationships of Plecoptera.•Plecoptera is sister-group to a clade composed of all other Polyneoptera orders.•The divergence between ...the suborders was the Pangaea origin.
Phylogenetic analysis based on mitochondrial genomic data from 25 stonefly species recovered a well-supported tree resolving higher-level relationships within Plecoptera (stoneflies). The monophyly of both currently recognized suborders was strongly supported, concordant with previous molecular analyses of Plecoptera. The southern hemisphere suborder Antarctoperlaria formed two clades: Eustheniidae + Diamphipnoidae and Austroperlidae + Gripopterygidae; consistent with relationships proposed based on morphology. The largely northern hemisphere suborder Arctoperlaria also divided into two groups, Euholognatha and Systellognatha, each composed of the five families traditionally assigned to each infraorder (the placement Scopuridae by mt genome data remains untested at this time). Within Euholognatha, strong support for the clade Nemouridae + Notonemouridae confirmed the northern origin of the currently southern hemisphere restricted Notonemouridae. Other family level relationships within the Arctoperlaria differ from those recovered by previous morphology and molecular based analyses. A fossil-calibrated divergence estimation suggests the formation of two suborders dates back to the Jurassic (181 Ma), with subsequent diversification of most stonefly families during the Cretaceous. This result confirms the hypothesis that initial divergence between the suborders was driven by the breakup of the supercontinent Pangaea into Laurasia and Gondwanaland (commencing 200 Ma and complete by 150 Ma).
Mayflies, stoneflies and caddisflies (Ephemeroptera, Plecoptera and Trichoptera) are prominent representatives of aquatic macroinvertebrates, commonly used as indicator organisms for water quality ...and ecosystem assessments. However, unambiguous morphological identification of EPT species, especially their immature life stages, is a challenging, yet fundamental task. A comprehensive DNA barcode library based upon taxonomically well‐curated specimens is needed to overcome the problematic identification. Once available, this library will support the implementation of fast, cost‐efficient and reliable DNA‐based identifications and assessments of ecological status. This study represents a major step towards a DNA barcode reference library as it covers for two‐thirds of Germany's EPT species including 2,613 individuals belonging to 363 identified species. As such, it provides coverage for 38 of 44 families (86%) and practically all major bioindicator species. DNA barcode compliant sequences (≥500 bp) were recovered from 98.74% of the analysed specimens. Whereas most species (325, i.e., 89.53%) were unambiguously assigned to a single Barcode Index Number (BIN) by its COI sequence, 38 species (18 Ephemeroptera, nine Plecoptera and 11 Trichoptera) were assigned to a total of 89 BINs. Most of these additional BINs formed nearest neighbour clusters, reflecting the discrimination of geographical subclades of a currently recognized species. BIN sharing was uncommon, involving only two species pairs of Ephemeroptera. Interestingly, both maximum pairwise and nearest neighbour distances were substantially higher for Ephemeroptera compared to Plecoptera and Trichoptera, possibly indicating older speciation events, stronger positive selection or faster rate of molecular evolution.
