Identification of clonal somatic mutations in DNA sequencing data from single cells
Joanna Hård1,5, Marie Kindblom1,5, Ezeddin Al Hakim1,5, Jeff E. Mold1,5, Marta Paterlini1, Pedro Reu1, Åsa Björklund2, Jens Lagergren3, Erik Borgström3, Kanar Alkass4, Henrik Druid4, Jonas Frisén1
1Department of Cell and Molecular Biology, Karolinska Institutet, Stockholm, Sweden; 2Science for life laboratory, Uppsala, Sweden; 3Science for life laboratory, KTH Royal Institute of Technology, Solna, Sweden; 4Department of Forensic Medicine, Karolinska Institutet, Stockholm, Sweden
Somatic mutations acquired by single cells can be used to reconstruct cell lineage trees. Unfortunately, variant calling in DNA sequencing data from single cells is challenging due allelic dropout and amplification bias in whole genome amplified genomic DNA. Specifically, in samples exhibiting high allelic dropout rates, it is not possible to conclude absence of a mutation based on absence of signal. To circumvent these problems we developed Conbase, a method for calling clonal somatic mutations in DNA sequencing data from whole genome amplified single cells. Somatic mutations are detected by linking each putative variant to adjacent, donor-specific heterozygous single nucleotide polymorphisms. This greatly reduced the contribution of sequencing and amplification errors, expediting the identification of variants that can be used to define phylogenetic relationships at a single cell level, and allowing us to confirm absence of mutations despite high rates of allelic dropout. We have validated Conbase by examining clonal somatic mutations in cultured primary human cells with a known cellular relationship. We are now further expanding our studies to analyze somatic mutational burden in a variety of postmortem tissues, in order to address questions about the origins and precise turnover dynamics of individual cells in human beings.