Counting and correcting errors within unique molecular identifiers to generate absolute numbers of sequencing molecules [RNA-seq]

Unique Molecular Identifiers (UMIs) are random oligonucleotide barcodes sequences? that are critical for the removal of PCR amplification biases within both bulk and single-cell sequencing experiments. However, the impact that PCR and sequencing errors have on the accuracy of generating absolute counts of RNA molecules is underappreciated. We demonstrate that PCR errors and not sequencing errors are the main source of inaccuracy in sequencing data and that the use of UMIs synthesized with homotrimeric nucleoside building blocks provides a solution to pinpoint and remove errors, allowing absolute counting of sequenced molecules. Overall design: In order to evaluate the accuracy of our homotrimer barcode assignment we processed mixed mixed mouse and human mRNA (at ratio of 50:50) and then performed bulk RNA sequencing.