Working phylogenetic hypothesis. Most relevant clades have sensitivity plots showing their prevalence among different scenarios.

Pipeline efficiency and gene prevalence after annotations.

Normalized match split distances among selected cladograms. (*) Bayesian inference was only performed for complete annotated datasets using translation-based alignment.

Without gene annotation, the alignment of the entire polyprotein might lead to spurious alignments in which nucleotides from different genes are put together into the same character.

A new annotation pipeline for Flaviviridae genomes, including: (i) prediction of putative protein coding sequences with GeneWise using reference protein sequences from UniProt and NCBI; (ii) validation of best matches using TransDecoder, BlastP, and hmmscan; (iii) pooling orthologous loci for translation alignment using PAM250; (iv) removal of outliers through the Tukey method. We joined genes and aligned them with different algorithms (ClustalW, Mafft, Muscle, Geneious translation-based).