DNA-stabilized silver nanoclusters (Ag_N-DNAs) have sequence-tuned compositions and fluorescence colors. High-throughput experiments together with supervised machine learning models have recently enabled design of DNA templates that select for Ag_N-DNA properties, including near-infrared (NIR) emission that holds promise for deep tissue bioimaging. However, these existing models do not enable simultaneous selection of multiple Ag_N-DNA properties, and require significant expert input for feature engineering and class definitions. This work presents a model for multiobjective, continuous-property design of Ag_N-DNAs with automatic feature extraction, based on variational autoencoders (VAEs). This model is generative, i.e., it learns both the forward mapping from DNA sequence to Ag_N-DNA properties and the inverse mapping from properties to sequence, and is trained on an experimental data set of DNA sequences paired with Ag_N-DNA fluorescence properties. Experimental testing shows that the model enables effective design of Ag_N-DNA emission, including bright NIR Ag_N-DNAs with 4-fold greater abundance compared to training data. In addition, Shapley analysis is employed to discern learned nucleobase patterns that correspond to fluorescence color and brightness. This generative model can be adapted for a range of biomolecular systems with sequence-dependent properties, enabling precise design of emerging biomolecular nanomaterials.

Objectives

This study sought to determine the relationship between long QT syndrome (LQTS) subtype (LTQ1, LTQ2, LTQ3) and postnatal cardiac events (CEs).

Background

LQTS presenting with 2:1 atrioventricular block or torsades de pointes in the fetus and/or neonate has been associated with risk for major CEs, but overall outcomes and predictors remain unknown.

Methods

A retrospective study involving 25 international centers evaluated the course of fetuses/newborns diagnosed with congenital LQTS and either 2:1 atrioventricular block or torsades de pointes. The primary outcomes were age at first CE after dismissal from the newborn hospitalization and death and/or cardiac transplantation during follow-up. CE was defined as aborted cardiac arrest, appropriate shock from implantable cardioverter-defibrillator, or sudden cardiac death.

Results

A total of 84 fetuses and/or neonates were identified with LQTS (12 as LQT1, 35 as LQT2, 37 as LQT3). Median gestational age at delivery was 37 weeks (interquartile range: 35 to 39 weeks) and age at hospital discharge was 3 weeks (interquartile range: 2 to 5 weeks). Fetal demise occurred in 2 and pre-discharge death in 1. Over a median of 5.2 years, there were 1 LQT1, 3 LQT2, and 23 LQT3 CEs (13 aborted cardiac arrests, 5 sudden cardiac deaths, and 9 appropriate shocks). One patient with LQT1 and 11 patients with LQT3 died or received cardiac transplant during follow-up. The only multivariate predictor of post-discharge CEs was LQT3 status (LQT3 vs. LQT2: hazard ratio: 8.4; 95% confidence interval: 2.6 to 38.9; p < 0.001), and LQT3, relative to LQT2, genotype predicted death and/or cardiac transplant (p < 0.001).

Conclusions

In this large multicenter study, fetuses and/or neonates with LQT3 but not those with LQT1 or LQT2 presenting with severe arrhythmias were at high risk of not only frequent, but lethal CEs.