A new open-source platform called UMIche improves accuracy in RNA-sequencing by fixing common counting errors. Published in Nature Communications, it provides a consistent framework for molecular data. Associate professor Adam Cribbs said, "By making UMI deduplication evidence-based rather than a default, UMIche sets a common standard for accurate molecule counting."
The open-source platform, called UMIche (pronounced "you-me-kay"), is described in a new study published in Nature Communications. It addresses a common challenge in RNA-sequencing, where errors during sequencing can lead to inaccurate measurements of gene expression.
Unique molecular identifiers (UMIs) are short barcodes added to each molecule before the molecules are copied through a process known as polymerase chain reaction (PCR). The UMI barcodes allow copies to be recognized and counted only once. However, errors introduced during PCR amplification and sequencing can alter these barcodes, potentially distorting the results.
Numerous mathematical and machine learning approaches have been developed for UMI error correction. But until UMIche, there has not been a consistent framework for testing and improving molecular counting methods.
The value of UMIche is to correct these errors and ensure accurate quantification.
Associate professor Adam Cribbs, co-author of the study from NDORMS, said, "By making UMI deduplication evidence-based rather than a default, UMIche sets a common standard for accurate molecule counting across the field."
