(Invited) Exploration of Short DNA Sequences Toward Complete Resolution of Single-Chirality SWCNTs

Abstract

Sequence-dependent interactions between DNA and single-wall carbon nanotubes (SWCNTs) have been shown to provide resolution for the atomic-structure-based sorting of DNA-wrapped SWCNTs by their helicity and handedness. Although significant progress has been made on DNA sequence selection for sorting SWCNTs, DNA sequence screening so far is still a high-cost, time-consuming and inefficient process. The success rate of finding recognition sequences is rather low (~10%), and it remains a grand challenge to predict recognition sequence patterns for sorting SWCNTs. Aiming to conquer this challenge, we have developed a highly efficient approach of finding recognition sequences by machine-learning guided experimental exploration of DNA sequence space. In this contribution, we present experimental data from short DNA sequences consisting of 5 and 6 nucleotides (referred to as 5-mers and 6-mers). Selective short DNA sequences demonstrate remarkable sorting resolution toward single-chirality SWCNTs, giving rise to more than 20 highly pure species. Since the short DNA sequences possess plain pattern combination and a relatively small number of distinct members of their population, we are able to explore the entire sequence space to address recognition pattern. Finally, we combine experimental investigation and machine learning to develop a highly efficient approach to differentiate recognition and nonrecognition sequences with success rate of > 90%.