Extended Data Fig. 5: Comparison of air eDNA long-read versus short-read shotgun sequencing.
a) Cumulative hits for species in each metazoan phylum, for long- and short-reads generated from the same air eDNA sample (1-week forest air April 2023), allowing for a direct comparison of sequencing output across the phyla. The y-axis lists the phyla, and the x-axis shows the total matches per phylum. b) Proportion of long- and short-reads stemming from the same air eDNA filtrate which map to any of 45 Florida or Ireland diverse eukaryote reference genomes with high quality (MAPQ ≥ 30). Log-log scatter plot of the fraction of species-uniquely mapped short-reads (Illumina, X-axis) or long-reads (ONT, Y-axis). Lines represent log-log linear regression, gray shading 95% confidence interval. R2, coefficient of determination. Left: correlations for the 1-week Florida forest air April 2023 sample. Right: correlations for the 8-day Wicklow mountain air July 2023 sample. Left: relative species abundance of forest air eDNA, as measured by short-read (n = 44,319 out of 6 Million subsampled reads) and long-read (n = 73,772 out of 6,509,032 reads) sequencing. Right: relative species abundance of mountain air eDNA, as measured by short-read (n = 92,421 out of 6 Million subsampled reads) and long-read (n = 160,104 out of 1,424,249 reads) sequencing. Overlaid species silhouette images from https://www.phylopic.org. c) Genus call overlaps between dual long- and short-read sequencing of the same air eDNA samples: 8-day Dublin city air July 2023 (left) and 3-week Florida forest air November 2023 (right). Includes all domain-of-life genera, except deuterostomes. CZ ID cloud-based metagenomics analysis.
