CaltechAUTHORS
  A Caltech Library Service

Scaling up molecular pattern recognition with DNA-based winner-take-all neural networks

Cherry, Kevin M. and Qian, Lulu (2018) Scaling up molecular pattern recognition with DNA-based winner-take-all neural networks. Nature, 559 (7714). pp. 370-376. ISSN 0028-0836. http://resolver.caltech.edu/CaltechAUTHORS:20180214-162530340

[img] Image (JPEG) (Extended Data Fig. 1: DNA implementation of winner-take-all neural networks) - Supplemental Material
See Usage Policy.

79Kb
[img] Image (JPEG) (Extended Data Fig. 2: Seesaw circuit implementation of winner-take-all neural networks) - Supplemental Material
See Usage Policy.

70Kb
[img] Image (JPEG) (Extended Data Fig. 3: Experimental characterization of winner-take-all DNA neural networks) - Supplemental Material
See Usage Policy.

155Kb
[img] Image (JPEG) (Extended Data Fig. 4: A winner-take-all DNA neural network that recognizes 9-bit patterns as ‘L’ or ‘T’) - Supplemental Material
See Usage Policy.

131Kb
[img] Image (JPEG) (Extended Data Fig. 5: A winner-take-all DNA neural network that recognizes 100-bit patterns as one of two handwritten digits) - Supplemental Material
See Usage Policy.

149Kb
[img] Image (JPEG) (Extended Data Fig. 6: Three-species winner-take-all behaviour and rate measurements for selecting DNA sequences in winner-take-all reaction pathways) - Supplemental Material
See Usage Policy.

151Kb
[img] Image (JPEG) (Extended Data Fig. 7: A winner-take-all DNA neural network that recognizes 100-bit patterns as one of three handwritten digits) - Supplemental Material
See Usage Policy.

199Kb
[img] Image (JPEG) (Extended Data Fig. 8: Workflow of the winner-take-all compiler) - Supplemental Material
See Usage Policy.

46Kb
[img] Image (JPEG) (Extended Data Fig. 9: Size and performance analysis of logic circuits for pattern recognition) - Supplemental Material
See Usage Policy.

111Kb
[img] Image (JPEG) (Extended Data Fig. 10: Species and their initial concentrations in all neural networks that recognize 100-bit patterns) - Supplemental Material
See Usage Policy.

112Kb
[img] MS Excel (Supplementary Table 1: DNA sequences) - Supplemental Material
See Usage Policy.

23Kb
[img] MS Excel (Supplementary Table 2: Weights and inputs) - Supplemental Material
See Usage Policy.

70Kb
[img] MS Excel (Supplementary Table 3: Experimental protocols) - Supplemental Material
See Usage Policy.

475Kb
[img] MS Excel (Source Data for Figure 2) - Supplemental Material
See Usage Policy.

579Kb
[img] MS Excel (Source Data for Figure 3) - Supplemental Material
See Usage Policy.

167Kb
[img] MS Excel (Source Data for Figure 4) - Supplemental Material
See Usage Policy.

1860Kb
[img] MS Excel (Source Data for Extended Data Figure 3) - Supplemental Material
See Usage Policy.

618Kb
[img] MS Excel (Source Data for Extended Data Figure 4) - Supplemental Material
See Usage Policy.

180Kb
[img] MS Excel (Source Data for Extended Data Figure 5) - Supplemental Material
See Usage Policy.

84Kb
[img] MS Excel (Source Data for Extended Data Figure 6) - Supplemental Material
See Usage Policy.

148Kb
[img] MS Excel (Source Data for Extended Data Figure 7) - Supplemental Material
See Usage Policy.

537Kb
[img] PDF - Published Version
See Usage Policy.

29Mb

Use this Persistent URL to link to this item: http://resolver.caltech.edu/CaltechAUTHORS:20180214-162530340

Abstract

From bacteria following simple chemical gradients to the brain distinguishing complex odour information, the ability to recognize molecular patterns is essential for biological organisms. This type of information-processing function has been implemented using DNA-based neural networks, but has been limited to the recognition of a set of no more than four patterns, each composed of four distinct DNA molecules. Winner-take-all computation has been suggested as a potential strategy for enhancing the capability of DNA-based neural networks. Compared to the linear-threshold circuits and Hopfield networks used previously, winner-take-all circuits are computationally more powerful, allow simpler molecular implementation and are not constrained by the number of patterns and their complexity, so both a large number of simple patterns and a small number of complex patterns can be recognized. Here we report a systematic implementation of winner-take-all neural networks based on DNA-strand-displacement reactions. We use a previously developed seesaw DNA gate motif, extended to include a simple and robust component that facilitates the cooperative hybridization that is involved in the process of selecting a ‘winner’. We show that with this extended seesaw motif DNA-based neural networks can classify patterns into up to nine categories. Each of these patterns consists of 20 distinct DNA molecules chosen from the set of 100 that represents the 100 bits in 10 × 10 patterns, with the 20 DNA molecules selected tracing one of the handwritten digits ‘1’ to ‘9’. The network successfully classified test patterns with up to 30 of the 100 bits flipped relative to the digit patterns ‘remembered’ during training, suggesting that molecular circuits can robustly accomplish the sophisticated task of classifying highly complex and noisy information on the basis of similarity to a memory.


Item Type:Article
Related URLs:
URLURL TypeDescription
https://doi.org/10.1038/s41586-018-0289-6DOIArticle
https://rdcu.be/2vscPublisherFree ReadCube access
ORCID:
AuthorORCID
Qian, Lulu0000-0003-4115-2409
Additional Information:© 2018 Macmillan Publishers Limited, part of Springer Nature. Received: 30 October 2017; Accepted: 18 April 2018; Published online 4 July 2018. We thank R. M. Murray for sharing an acoustic liquid-handling robot. We thank C. Thachuk and E. Winfree for discussions and suggestions. K.M.C. was supported by a NSF Graduate Research Fellowship. L.Q. was supported by a Career Award at the Scientific Interface from the Burroughs Wellcome Fund (1010684), a Faculty Early Career Development Award from NSF (1351081), and the Shurl and Kay Curci Foundation. Reviewer information: Nature thanks R. Schulman and the other anonymous reviewer(s) for their contribution to the peer review of this work. Author Contributions: K.M.C. developed the model, designed and performed the experiments, and analysed the data; K.M.C. and L.Q. wrote the manuscript; L.Q. initiated and guided the project. Data availability: All data that support the findings of this study are included in the manuscript and its Extended Data. Source Data for Figs. 2–4 and Extended Data Figs. 3–7 are provided with the online version of the paper. The authors declare no competing interests.
Funders:
Funding AgencyGrant Number
NSF Graduate Research FellowshipUNSPECIFIED
Burroughs Wellcome Fund1010684
NSFCCF-1351081
Shurl and Kay Curci FoundationUNSPECIFIED
Record Number:CaltechAUTHORS:20180214-162530340
Persistent URL:http://resolver.caltech.edu/CaltechAUTHORS:20180214-162530340
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:84840
Collection:CaltechAUTHORS
Deposited By: George Porter
Deposited On:05 Jul 2018 17:11
Last Modified:26 Jul 2018 20:50

Repository Staff Only: item control page