Fig. 4: Determination of the screening efficiency for single-cell platform-based BsAb functional screening, using HEK293_LP^CD19/Blin-spiked cell samples.

a Scatterplot showing the observed recovery rate versus the abundance of HEK293_LP^CD19/Blin spiked in bulk negative cells (one million HEK_LP^CD19). The recovery rate is defined as the isolated copy number divided by the total copy number initially spiked into a sample. Each data point (“dot”) is derived from the mix of \(n=3\) analytical replicates of spiked samples. The solid red line is derived from linear regression with one standard deviation bounds (dotted lines). b Predicted recovery rate based on a linear regression model and its 90% confidence interval (R² = 0.34; p-value = 0.0002; statistical assumptions largely fulfilled, e.g., stable residual variance). c Prediction of screening efficiency (“P”), defined as the probability of discovering ≥1 copy of a positive clone from a cell pool. This model is based on a binomial probability, where the Bernoulli probability “p” refers to the predicted recovery rate in the foregoing regression model, corresponding to a selected confidence lower bound (90%-solid red line or 95%-solid blue line).