Using B-allele frequency plots, we now screen our genome data for contamination as well as aneuploidy

Phylogenetic trees were wrong for in silico contaminated data from 2 different fungal species - S. cerevisiae and A. fumigatus - and using multiple tools (incl. NJ and ML approaches).

Surprisingly, even 5-10% contamination can lead to incorrect tree topologies.

Phylogenetic tools usually treat heterozygous sites as missing data, so we imagined this would not be a problem. By contaminating genome data in silico, we checked.

Unsurprisingly, contamination raised levels of “heterozygous” base calls, and strains appeared genetically admixed.