Eukaryotic gene prediction on metagenomic scaffolds. (A) Gene predictions for nine diverse eukaryotic organisms including fungi, a Metazoa, a Stramenopile, an Archaeplastida, and a Rhizaria. Columns labeled “R” refer to reference gene sets, whereas M columns refer to gene sets predicted without transcript or close homology evidence. The top panel displays the proportion of total genes either overlapping (shared) or not overlapping (unshared) a gene model from the other respective gene set for a given genome. The bottom panel is an analysis of presence or absence of single-copy genes in each gene set as determined by BUSCO using the eukaryota_odb9 lineage set. (B) Proportion of a soil metagenome spiked with the genome of Choanephora cucurbitarum predicted to be either noneukaryotic, eukaryotic and belonging to the Choanephora, or predicted to be eukaryotic but has homology to prokaryotic sequences. (C) BUSCO analysis of the binned Choanephora cucurbitarum genome with protein sets from (left to right) the reference protein set, trained MAKER2 output, and whole metagenome MetaProdigal output.
