Extended Data Fig. 2: Sequence tags enable lineage-tracing of edited cells.

a, Expected and observed sequence tags targeting the E-box at the SHMT2 promoter. Sequence logos for expected distribution of flanking mutations were generated on the basis of the mutation strategy used for generating the library of HDR templates: each non-consensus base is represented with a probability of 8%. The observed sequence logos were generated from all sequence tags observed in the ChIP input sample for targeted SHMT2 E-box locus. The sequence tags within the HDR templates with original E-box sequence are shown on the left (from 8442 recovered tags), and with the mutated TATTTA sequence on the right (from 9128 recovered tags). ChIP input is collected 48 h after transfection and thus represents the baseline of transfected HDR templates in a single experiment. Note that the consensus bases are taller in the logo representing the expected sequences compared to the observed ones, since expected sequences also include the variants without any flanking mutations (that are inevitably generated in this mutation strategy), whereas the observed logos are generated for the sequence tags with at least one mutation. b, Precision editing results for the effect of E-box mutation at the RPL23 gene promoter on fitness of HAP1 cells separately for each cell lineage pair with a sequence tag having exactly one flanking mutation (x-axis on logarithmic scale, see also Fig. 1d). Read count ratios for mutated vs. original sequence are shown at two time points, day 8 (grey) and day 2 (pink).