A General Approach for Machine Learning-Aided Design of DNA-Stabilized Silver Clusters

DNA-templated silver clusters (AgN-DNA) are known to exhibit a wide range of fluorescence colors for different choices of the DNA template sequence. While these clusters are promising biosensors and biomarkers, rational design of AgN-DNA is challenged by the huge space of possible DNA template sequences. Recent work used machine learning methods trained on experimental data to design new DNA templates that select for AgN-DNA color, for the specific case of 10-base DNA oligomers. An important open question is whether such a design process developed for a specific biopolymer template length is applicable at other lengths, with different numbers and diversities of cluster nucleation sites. Here, we develop a flexible design approach that builds on color-correlated DNA base motifs learned from data on more than 2,000 10-base DNA oligomers. We test this motif-based design for templates ranging from 8 bases to 16 bases long, spanning a factor of > 65000 in the size of sequence space. Designed strands of all len...