Fig. 1

a An overview conceptualising the problem with current gene annotation methods and the stages Panaroo uses to correct for annotation errors. b Expanded specific stages in the process. (i) Contamination appears in the graph as poorly supported components. In the default mode, Panaroo removes contamination by recursively removing poorly supported nodes of degree 1. (ii) Genes are often mis-annotated near contig breaks [19]. Panaroo corrects such mis-annotations by recursively removing poorly supported nodes of degree 1. (iii) Panaroo corrects cases where the same DNA sequence has been translated in multiple reading frames into a single gene by clustering concomitant genes at the DNA level. (iv) Panaroo uses context and a lower clustering threshold to combine diverse gene families into a single gene. (v) Annotation algorithms may predict a gene in some but not all samples, even when the samples share exactly the same DNA sequence. Panaroo finds missing genes by searching for the gene sequence in the surrounding DNA