Extended Data Fig. 4: Simulating proliferation of cells with the scalable CRISPR-barcode system.
From: Deep distributed computing to reconstruct extremely large lineage trees
a, The scGESTALT lineage tracing datasets of zebrafish brain development. BBI scores were obtained for branches that had more than three leaves associated to each of the two descending edges and a total of more than ten leaves associated to both of the two descending edges. b, c, Estimation of the insertion and deletion event probabilities per barcode array per generation in the scGESTALT dataset. The relative insertion and deletion event probabilities across each barcode array position and the probability distributions of insertion and deletion sizes were modeled using those observed in the scGESTALT dataset. The probabilities of the insertion and deletion events per generation for the production of 4,000 cells were fitted to the average total insertion and deletion lengths per barcode array observed in the scGESTALT dataset, respectively. d, Median fractions of base substitution per nucleotide position per generation observed for different secondary scaling factors for base editing in the simulation of producing 4,000 cells. The blue shading represents the 5 to 95 percentile range. e, f, Tree recovery scores for RapidNJ and FRACTALized RapidNJ in reconstructing the lineages of 4,000 cells simulated based on different secondary scaling factors for base editing.
