In their manuscript “Single molecule counting and assessment of random molecular tagging errors with transposable giga-scale error-correcting barcodes”, Lau and Ji present the design and synthesis of a library of 262,144 different indexed oligonucleotides adapters, which they call EXBs, for “Exponentially-expanded Barcodes”. Each EXB is a combination of three 6-mer barcodes taken from a set of 64 sequences chosen for having a high edit distance between each others. To be used as an alternative to random unique molecular identifiers (UMIs) in quantitative sequencing, the EXBs are be combined in more than 69 billions of pairs, introduced with the Tn5 transposase.

The authors compare the performance of EXBs and UMIs by preparing transcriptome libraries with adapter sequences containing both of them, and show that spurious differential expression is observed when using UMIs without correction. Incidentally, this dataset also provides precise information about the error profile in UMIs. EXBs are a novel way to introduce molecular barcodes in double-stranded DNA molecules.

Using the EXBs requires to read 142 nt of linker sequences, which considerably add to the cost on current platforms such as Illumina's HiSeq or NovaSeq. Thus they may be more useful with long-read sequencers such as the ones from Oxford Nanopore Technologies.