Figure 1. Haplotype-aware discovery of SVs in single cells.
(a) Overview of the Strand-seq protocol used to preserve strand-orientation and homolog (haplotype) identity. BrdU (Bromodeoxyuridine) is incorporated into dividing cells, followed by removal of the BrdU-containing strands (dashed line) through nicking, and short read sequencing of the remaining (template, solid line) strand21. W, Watson strand (orange); C, Crick (blue); H, haplotype. Right panel: haplotagging approach that assigns individual Strand-seq reads to either haplotype 1 (H1) or H2. Red lollipops mark reads assigned to H1 based on overlapping SNPs; blue lollipops mark H2 reads. From this, three data layers are considered: 1. the total number of reads in a binned region are measured to calculate the ‘Depth’ layer, 2. the relative proportion of W and C reads are measured to calculate the ‘Strand’ layer, and 3. the number of W and C reads assigned to H1 or H2 are used to calculate the ‘Phase’ layer. (b) Scheme depicting how strands segregate during mitosis to reveal SVs in single cells. Del, deletion; Inv, inversion; Tr, translocation. Segments of derivative chromosomes share the same strand label during DNA replication and co-segregate. H1/H2 and h1/h2 designate haplotypes 1 and 2 for two different chromosomes. (c) scTRIP exploits read depth, strand ratio, and chromosome-length haplotype phase as data layers. Haplotype phase is assessed in a strand-aware fashion, with phased W reads shown as lollipops on left of ideogram and phased C reads shown to right (using the same haplotype colors as in (a)). An example “reference” state is shown, which contains 2N read depth, equal proportion of W:C reads and both haplotypes. Panels (d-f) depict diagnostic footprints for chromosomes where both haplotypes are labeled on different strands (‘WC/CW chromosomes’). Our framework also detects and scores equivalent footprints on CC and WW chromosomes (see Table S1). (d) Deletion (Del), detected as losses in read depth affecting a single haplotype, combined with unaltered read orientation. Duplication (Dup), detected as a haplotype-specific gain in depth with unaltered read orientation. (e) Balanced inversion (Inv), identified as haplotype-phased read orientation ‘flips’ with unaltered depth. Inverted duplication (InvDup), characterized by inverted reads detected for one haplotype coinciding with a read depth gain of the same haplotype. (f) Ploidy alterations can be detected as departures from diploid W and C segregation ratios (see also Table S2). (g) Balanced translocation show correlated template strand switches affecting the same paired genomic regions in cells harboring the SV.