Additional file 6: of Assessing characteristics of RNA amplification methods for single cell RNA sequencing

Gene detection logistic regression fit and validation. Model was fit using randomly selected 90% of 10 pg. data, excluding 17 large influence genes. Fit was evaluated on the remaining 10% of the data. Fit was also evaluated on sequence data that was in silico truncated to 50 base pair single end (“Truncated”), ERCC read counts (“ERCC”), and 100 pg. dilution replicates (“100 pg.”). AUC (area under receiver operating characteristic curve) reported as mean values ± 2 Sd. calculated over 10,000 bootstrap samples. AUC (molecules) predicts detection based on number of input molecules alone. See Methods for further details. (XLS 29 kb)