Fig. 1

The pooled amplification paradox. Illustration of a seemingly paradoxical reversal in gene abundance estimates arising from incomplete sampling of a cDNA library generated after a differential amplification of two genes. Consider a situation where more molecules have been captured for Gene 1 (8) than Gene 2 (6). Despite the correction for amplification bias with unique molecular identifiers, the increased amplification of Gene 2 relative to Gene 1, coupled with the incomplete sequencing of the cDNA library, results in a lower copy number estimate of Gene 1 (4) in comparison to Gene 2 (5). Correction of abundance estimates due to unseen molecules using a histogram of counts per UMI improves the quantification, reversing the relative amount of Gene 1 and Gene 2