systematic Archives

Neural network training of varKodes for species identification. (A) Effect of k-mer length and data amount used to produce varKodes on validation accuracy. Longer k-mers increase accuracy when more data are used. Mixing varKodes subsampled from different amounts of data improves accuracy. Box with dashed line (k-mer length = 7) strikes a good balance between model accuracy and amount of required data. (B) Validation accuracy improves with increased number of training samples per species, but even 3–4 samples are sufficient in most cases for achieving high accuracy. Each solid line represents one sample, colored by DNA quality (i.e., variation in base pair frequencies). Higher rank indicates better quality. Dashed lines represent averages across all samples. -- ecoevorxiv.org

AI - Data - Apps - Cybernetics, Analog Studies, Away Teams & Field Reports, Biosignatures & Paleobiology, Fossils & Paleontology, Mapping, Geodesy, Cartography, Bathymetry, Nomenclature, Systematics, Natural History, Status Report

Getting Ready For Offworld Astrobiology Expeditions: A Universal DNA Barcode For The Tree Of Life

by Keith CowingFebruary 7, 2024February 7, 2024

Editor’s note: According to this study “Species identification using DNA barcodes has revolutionized biodiversity sciences and society at large.“ When we start to explore inhabited worlds we’re going to have […]

AI - Data - Apps - Cybernetics, Analog Studies, Arctic / Antarctic / Alpine, Atmospheres, Climate, Weather, Biogeochemical Cycles & Geobiology, Commercial & Private Sector, Extremeophiles and Extreme Environments, Genomics, Proteomics, Bioinformatics, Nanotechnology & SynBio, Nomenclature, Systematics, Natural History, Tricorders & Sensors

These Away Teams Are Building The Largest Knowledge Graph Of Life On Earth

by Keith CowingDecember 4, 2023September 18, 2025

Editor’s note: As we search for life on other worlds via orbital and surface surveys, it is useful to use Earth and its diverse biota on land, sea, and air, […]