Fig. 2

Dramatic effect of cross contaminations on reconstructing ctenophore relationships using a phylogenomic dataset. a Cross contamination network for dataset A as reconstructed by CroCo. Nodes and links represent transcriptomes and cross contaminations, respectively. Only transcripts strictly categorised as cross contaminants are taken into account here. Node sizes are proportional to the number of times the node is the source of cross contamination, and node colours represent the percentage of contaminated transcripts in the transcriptome. For clarity, weak links, defined as less than 2% of the strongest link in the network, are not shown. b, c Ctenophore phylogenetic relationships reconstructed with 114 genes using (b) untreated transcriptomes and (c) transcriptomes cleaned using CroCo (see details in Methods). In (b) the placement of lineages highlighted in orange disrupts the monophyly of the clade ‘Lobata’ (here represented by Mnemiopsis leidyi and Bolinopsis infundibulum). With the cleaned dataset (c), the same lineages, in blue, are placed in agreement with recent studies of ctenophore relationships [16, 31,32,33]. The two dotted arrows and their corresponding numbers indicate two major cross contamination events that can be observed on the cross contamination network (a)