DNA-stabilized silver clusters (AgN-DNAs) consist of ~10-30 silver atoms encapsulated within DNA oligomers. These fluorescent clusters have sequence-tuned colors: by choice of nucleobase sequence, AgN-DNA fluorescence emission can be tuned from about 450 nm up to at least 1000 nm [1]. AgN-DNAs are promising fluorophores and sensors [2], but their rational design remains challenging due to the huge space of possible DNA template sequences. To discern how DNA sequence selects AgN size and fluorescence spectrum, we employ high-throughput experiments and tools from machine learning and data mining. Our past work showed that machine learning classifiers trained on experimental data in the limited case of 10-base DNA oligomers can be used to design new 10-base DNA templates which are selective of AgN-DNA fluorescence emission color [3]. However, because most AgN-DNA researchers use longer oligomers, it is critical to determine whether this approach is predictive at longer oligomer lengths with greater numbers and configurations of cluster nucleation sites. This question relates to a broader question within materials informatics: when are models learned on limited training data sets relevant within an expanded input parameter space, and how can we use experimental data sets for effective informatics approaches to materials design? Here, we present a machine learning classification approach for AgN-DNA prediction which is general to multiple DNA oligomer lengths. By engineering features which quantify the occurences of color-correlated DNA base "motifs" learned from data on &gt; 2,000 10-base DNA oligomers, a set of supervised classifiers is trained to predict AgN-DNA color agnostic to sequence length [4]. We test this motif-based design for templates ranging from 8 bases to 16 bases long, for which the sizes of the sequence spaces differ by nearly 5 orders of magnitude. Experimental data show that designed strands of all lengths are selective for AgN-DNA color in the target wavelength band of 600-660 nm, strongly suggesting that color-selective motifs learned for one template length generalize to other lengths. Thus, a motif-based design approach may be broadly suitable for future AgN-DNA applications.

Discerning the color-sequence connection of fluorescent DNA-stabilized silver clusters by machine learning

ACS Spring Meeting 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.