Extended Data Fig. 10: Simulations of heating and cooling two hairpins in catalyzed hairpin assembly.
From: Heat-rechargeable computation in DNA logic circuits and neural networks
a, Reactions involved in the simulations, where n = 6 is the number of nucleotides in each domain shown in Fig. S10 in the Supplementary Information of Yin et al., 200844. b-d, If bimolecular hybridization rate does not depend on the length of the hybridization domain, the hairpin reset success rate is expected to be 85.1% for 20 nM hairpins heated to 95 °C for 5 minutes and cooled to 20 °C in 1 minute (b), 53.6% for 100 nM hairpins cooled in 1 minute (c), and 42.3% for 20 nM hairpins cooled in 10 minutes (d). e-g, If bimolecular hybridization rate depends on the length of the hybridization domain, the hairpin reset success rate is expected to be 42.5% for 20 nM hairpins heated for 5 minutes and cooled in 1 minute (e), 13.6% for 100 nM hairpins cooled in 1 minute (f), and 10.2% for 20 nM hairpins cooled in 10 minutes (g). The poor reset performance can be explained by the substantially longer hybridization domain in the undesired product AB compared to the desired hairpins A and B. A higher concentration results in both higher melting temperature of the undesired product and smaller kinetics difference between the desired and undesired reactions, worsening the reset performance. Length-dependent bimolecular hybridization rate further favors the undesired reaction. A longer cooling time allows for more undesired product AB to form before desired hairpin B and then A starts to form, resulting in worse reset. These observations are consistent with the principles discussed in Extended Data Fig. 3.
