As part of our NGS pipeline propagating uncertainties quantitatively, a novel module had been developed for the fusion of multiple callers. Such precision can be crucial for extremely high coverage, e.g. in oncology, immunogenomics or in mtDNA analysis. This deep fusion is performed in an early phase using raw descriptors from the individual callers and using heuristic, optimization and Bayesian averaging methods. For a related publication, see PMID: 26510841.