June 2025 · 6 Articles · Pages 201–285
June 2025
Open Access6 Articles · Pages 201–285
Cyanobacteria offer a promising biological platform for atmospheric carbon capture due to their high photosynthetic efficiency and genetic tractability. This review surveys recent advances in engineering cyanobacterial carbon concentrating mechanisms, RuBisCO variants, and carbon-sink metabolic pathways. We critically assess the scalability of these approaches and outline the key challenges remaining for industrial deployment of cyanobacteria-based carbon capture systems.
Telomeric and subtelomeric regions of the human genome are notoriously difficult to resolve with short-read sequencing technologies. Using Oxford Nanopore long-read sequencing across 50 human genomes, we catalog over 200 previously uncharacterized structural variants within telomeric regions, including insertions, deletions, and complex rearrangements. Several variants correlate with telomere length variation and may have implications for aging and cancer biology.
Elevation gradients provide natural laboratories for studying environmental drivers of microbial community composition. We characterized soil microbiomes across a 4,000-meter elevation gradient in the Peruvian Andes using 16S rRNA and ITS amplicon sequencing. Our results demonstrate distinct community turnover patterns for bacteria and fungi, with temperature and soil pH emerging as the dominant predictive variables. These data contribute to understanding how climate change may restructure high-altitude soil ecosystems.
Aberrant protein aggregation is a hallmark of amyotrophic lateral sclerosis (ALS), yet the kinetic mechanisms driving aggregate formation in vivo remain poorly understood. Using a combination of in vitro biophysical assays and live-cell imaging in SOD1 and TDP-43 mutant motor neuron models, we define the nucleation and elongation rates of pathological aggregates. Our kinetic framework identifies a critical concentration threshold for seeded aggregation that may inform therapeutic timing strategies.
We present a gradient-boosted machine learning framework trained on curated enzyme kinetics data from the BRENDA database that predicts Km and kcat values from amino acid sequence and substrate structure alone. The model achieves a mean absolute error within 0.5 log units for both parameters on held-out test sets, enabling rapid prioritization of enzyme variants for metabolic engineering applications.
Citizen science platforms generate vast quantities of spatiotemporal occurrence data for migratory species. We evaluated the accuracy of agent-based monarch butterfly migration models using over 120,000 georeferenced sightings from three citizen science databases spanning 2018–2024. Model predictions aligned closely with observed migration phenology and corridor usage, demonstrating that crowd-sourced data can serve as a reliable validation tool for complex ecological models.