In this project we sequenced ssDNA oligos generated by photolithography to evaluate the error rate parameter of the synthesis

Nucleic acid microarrays are the only tools that can supply very large oligonucleotide libraries, cornerstones of the nascent fields of de novo gene assembly and DNA data storage. Although the chemical synthesis of oligonucleotides is highly developed and robust, it is not error free, requiring sensitive applications to take this into account, designing methods to correct or compensate for errors or to select for high-fidelity oligomers. However, outside the realm of array manufacturers, little is known about the sources of errors and their extent. In this study, we look at the error rate of DNA libraries synthesized by photolithography and, via sequencing, dissect the proportion of deletion, insertion and substitution errors. We find that the deletion rate is governed by the photolysis yield and that it matches previous on-array measurements. We identify the most important substitution error and correlate it to phosphoramidite coupling. But besides synthetic failures originating from the coupling cycle, we uncover the role of imperfections and limitations related to optics as we highlight the importance of absorbing UV light to avoid internal reflections and as we chart the dependence of error rate on both position on the array and position within individual oligonucleotides. Being able to precisely quantify all types of errors will allow for optimal choice of fabrication parameters and array design, resulting in better DNA libraries with known error modes and distributions.