NANO COMPUTING DEVICE AND METHOD OF OPERATING NANO COMPUTING DEVICE
20210365798 · 2021-11-25
Inventors
Cpc classification
G11C13/02
PHYSICS
G06N3/002
PHYSICS
G06N3/06
PHYSICS
International classification
Abstract
A nano computing device includes: a nanoparticle memory including a first molecule bound so as to store a molecular input; a nanoparticle reporter including a second molecule bound so as to generate an output; and a nanoparticle floater including at least two third molecules and fourth molecules so as to be bound to one of the nanoparticle memory and the nanoparticle reporter based on the molecular input and an instruction molecule.
Claims
1. A nano computing device, comprising: a nanoparticle memory including a first molecule bound so as to store a molecular input; a nanoparticle reporter including a second molecule bound so as to generate an output; and a nanoparticle floater including at least two third molecules and fourth molecules so as to be bound to one of the nanoparticle memory and the nanoparticle reporter based on the molecular input and an instruction molecule.
2. The nano computing device of claim 1, wherein: the molecular input includes a molecular input indicating logic 1.
3. The nano computing device of claim 1, wherein: the instruction molecule includes at least one of a trap DNA that binds the nanoparticle floater to the nanoparticle memory and a report DNA that binds the nanoparticle floater to the nanoparticle reporter.
4. The nano computing device of claim 3, wherein: the trap DNA includes: a first trap DNA that binds the nanoparticle floater and the nanoparticle memory when a storage state of the nanoparticle memory indicates logic 0; and a second trap DNA that binds the nanoparticle floater and the nanoparticle memory when a storage state of the nanoparticle memory indicates logic 1.
5. The nano computing device of claim 4, wherein: the first trap DNA binds the first molecule and the third molecule when the storage state of the nanoparticle memory indicates logic 0.
6. The nano computing device of claim 4, wherein: the second trap DNA binds the molecular input bound to the first molecule and the third molecule when the storage state of the nanoparticle memory indicates logic 1.
7. The nano computing device of claim 1, wherein: a first reaction rate between the nanoparticle memory and the nanoparticle floater by a first instruction molecule among the instruction molecules is faster than a second reaction rate between the nanoparticle reporter and the nanoparticle floater by a second instruction molecule among the instruction molecules.
8. The nano computing device of claim 7, wherein: the first instruction molecule includes a trap DNA that binds the nanoparticle floater to the nanoparticle memory, the second instruction molecule includes a report DNA that binds the nanoparticle floater to the nanoparticle reporter, and a rate difference between the first reaction rate and the second reaction rate is adjusted according to at least one of a DNA hybridization domain configuring each of the first to fourth molecules, hybridization domains of the trap DNA and the report DNA, a concentration of the nanoparticle memory, a concentration of the nanoparticle reporter, a concentration of the nanoparticle floater, and a concentration of the trap DNA and the report DNA.
9. The nano computing device of claim 1, wherein: the nanoparticle memory and the nanoparticle reporter are fixed to a supported lipid bilayer membrane, and the nanoparticle floater is mobile on the supported lipid bilayer membrane.
10. A nano computing device, comprising: a plurality of molecular inputs; a plurality of nanoparticle memories configured to store the plurality of molecular inputs; a plurality of instruction molecules programmed for performing a logic operation; a plurality of nanoparticle reporters configured to determine an output; and a plurality of nanoparticle floaters bound to the plurality of nanoparticle memory or the plurality of nanoparticle reporters based on the molecular inputs and the instruction molecules.
11. The nano computing device of claim 10, wherein: the plurality of molecular inputs includes at least one of two types of a first molecular input and a second molecular input, and the plurality of nanoparticle memory includes a first nanoparticle memory and a second nanoparticle memory that store at least one of the first molecular input and the second molecular input.
12. The nano computing device of claim 11, wherein: the plurality of nanoparticle floaters is bound to the first nanoparticle memory by a first trap molecule among the plurality of instruction molecules, is bound to the second nanoparticle memory by a second trap molecule among the plurality of instruction molecules, and is bound to one of the plurality of nanoparticle reporters by a report molecule among the plurality of instruction molecules.
13. The nano computing device of claim 12, wherein: the first trap molecule binds one of the plurality of nanoparticle floaters to the first nanoparticle memory when a molecular input is not stored in the first nanoparticle memory, and the second trap molecule binds one of the plurality of nanoparticle floaters to the second nanoparticle memory when a molecular input is not stored in the second nanoparticle memory.
14. The nano computing device of claim 12, wherein: the first trap molecule binds one of the plurality of nanoparticle floaters to the first nanoparticle memory when a molecular input is not stored in the first nanoparticle memory, and the second trap molecule binds one of the plurality of nanoparticle floaters to the second nanoparticle memory when a second molecular input is stored in the second nanoparticle memory.
15. The nano computing device of claim 12, wherein: the first trap molecule binds one of the plurality of nanoparticle floaters to the first nanoparticle memory when a first molecular input is stored in the first nanoparticle memory, and the second trap molecule binds one of the plurality of nanoparticle floaters to the second nanoparticle memory when a second molecular input is stored in the second nanoparticle memory.
16. The nano computing device of claim 11, wherein: the plurality of nanoparticle floaters includes two types including a first nanoparticle floater and a second nanoparticle floater, the first nanoparticle floater is bound to the first nanoparticle memory by a first trap molecule among the plurality of instruction molecules and is bound to one of the plurality of nanoparticle reporters by a report molecule among the plurality of instruction molecules, and the second nanoparticle floater is bound to the first nanoparticle memory by a second trap molecule among the plurality of instruction molecules, is bound to the second nanoparticle memory by a third trap molecule among the plurality of instruction molecules, and is bound to another one of the plurality of nanoparticle reporters by a report molecule among the plurality of instruction molecules.
17. The nano computing device of claim 16, wherein: the first trap molecule binds the first nanoparticle floater and the first nanoparticle memory when a molecular input is not stored in the first nanoparticle memory, the second trap molecule binds the second nanoparticle floater and the first nanoparticle memory when a first molecular input is stored in the first nanoparticle memory, and the third trap molecule binds the second nanoparticle floater and the second nanoparticle memory when a molecular input is not stored in the second nanoparticle memory.
18. The nano computing device of claim 16, wherein: the first trap molecule binds the first nanoparticle floater and the first nanoparticle memory when a first molecular input is not stored in the first nanoparticle memory, the second trap molecule binds the second nanoparticle floater and the first nanoparticle memory when a molecular input is not stored in the first nanoparticle memory, and the third trap molecule binds the second nanoparticle floater and the second nanoparticle memory when a second molecular input is stored in the second nanoparticle memory.
19. The nano computing device of claim 11, wherein: the plurality of nanoparticle floaters includes two types including a first nanoparticle floater and a second nanoparticle floater, and the first nanoparticle floater is bound to the first nanoparticle memory by a first trap molecule among the plurality of instruction molecules, is bound to the second nanoparticle memory by a second trap molecule among the plurality of instruction molecules, and is bound to one of the plurality of nanoparticle reporters by a report molecule among the plurality of instruction molecules, and the second nanoparticle floater is bound to the first nanoparticle memory by a third trap molecule among the plurality of instruction molecules, is bound to the second nanoparticle memory by a fourth trap molecule among the plurality of instruction molecules, and is bound to another one of the plurality of nanoparticle reporters by a report molecule among the plurality of instruction molecules.
20. The nano computing device of claim 19, wherein: the first trap molecule binds the first nanoparticle floater and the first nanoparticle memory when a molecular input is not stored in the first nanoparticle memory, the second trap molecule binds the first nanoparticle floater and the second nanoparticle memory when a second molecular input is stored in the second nanoparticle memory, the third trap molecule binds the second nanoparticle floater and the first nanoparticle memory when a first molecular input is stored in the first nanoparticle memory, and the fourth trap molecule binds the second nanoparticle floater and the second nanoparticle memory when a molecular input is not stored in the second nanoparticle memory.
21. The nano computing device of claim 19, wherein: the first trap molecule binds the first nanoparticle floater and the first nanoparticle memory when a molecular input is not stored in the first nanoparticle memory, the second trap molecule binds the first nanoparticle floater and the second nanoparticle memory when a molecular input is not stored in the second nanoparticle memory, the third trap molecule binds the second nanoparticle floater and the first nanoparticle memory when a first molecular input is stored in the first nanoparticle memory, and the fourth trap molecule binds the second nanoparticle floater and the second nanoparticle memory when a second molecular input is stored in the second nanoparticle memory.
22. A nano computing device, comprising: an input layer including a plurality of nanoparticle memories; a hidden layer including a plurality of nanoparticle floaters; and an output layer including a plurality of nanoparticle reporters, wherein the nano computing device is programmed with a nanoparticle neural network including a plurality of instruction molecules corresponding to weights between the input layer and the hidden layer and between the hidden layer and the output layer.
23. The nano computing device of claim 22, wherein: the plurality of instruction molecules includes a plurality of trap molecules that controls binding between the plurality of nanoparticle memories and the plurality of nanoparticle floaters based on a storage state of the plurality of nanoparticle memories, and the type of the plurality of trap molecules is different according to the nanoparticle neural network.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0036]
[0037]
[0038]
[0039]
[0040]
[0041]
[0042]
[0043]
[0044]
[0045]
[0046]
[0047]
[0048]
[0049]
[0050]
[0051]
[0052]
[0053]
[0054]
[0055]
[0056]
DETAILED DESCRIPTION OF THE EMBODIMENTS
[0057] An exemplary embodiment according to the present invention implements a Nanoparticle-based Von Neumann Architecture (NVNA) by applying Von Neumann Architecture (VNA) to a lipid nanotablet which is capable of operating various programs without reconfiguring a computer. That is, in the exemplary embodiment, to create a stored-program device stored for programming in a molecular computing platform, the concept of memory that stores molecular information is incorporated to the VNA with nanoparticles. The stored molecular information is processed according to an input instruction code and the instruction code may be input by a user. The nanostructure (hardware) and the instruction code (software) are separated, so that the user is capable of performing multiple computational tasks only with updating of software without fabricating a new device every single time, thereby improving modularity and scalability in information processing in LNT.
[0058] According to the exemplary embodiment, arbitrary logic computing may be programmed several times on a single chip without reassembling. In the system, the nanoparticle on the lipid chip functions as hardware having characteristics of a memory, a processor, and an output device. Further, the system is software to provide molecular instructions for facile programming of logic circuits and uses DNA strands. Through the computing architecture, nanoparticles of one group form a feed-forward neural network, a perceptron, which may implement functionally complete Boolean logic operations. The NVNA according to the exemplary embodiment of the present invention provides a programmable, resettable, and scalable computing architecture and circuit board to form nanoparticle neural networks and make logical decisions.
[0059] Hereinafter, an exemplary embodiment disclosed the present specification will be described in detail with reference to the accompanying drawings, and the same or similar constituent element is denoted by the same and similar reference numeral, and a repeated description thereof will be omitted. Suffixes, “module” and “unit” for a constituent element used for the description below are given or mixed in consideration of only easiness of the writing of the specification, and the suffix itself does not have a discriminated meaning or role. Further, in describing the exemplary embodiment disclosed in the present disclosure, when it is determined that a detailed description relating to well-known functions or configurations may make the subject matter of the exemplary embodiment disclosed in the present disclosure unnecessarily ambiguous, the detailed description will be omitted. Further, the accompanying drawings are provided for helping to easily understand exemplary embodiments disclosed in the present specification, and the technical spirit disclosed in the present specification is not limited by the accompanying drawings, and it will be appreciated that the present invention includes all of the modifications, equivalent matters, and substitutes included in the spirit and the technical scope of the present invention.
[0060] Terms including an ordinary number, such as first and second, are used for describing various constituent elements, but the constituent elements are not limited by the terms. The terms are used only to discriminate one constituent element from another constituent element.
[0061] It should be understood that when one constituent element is referred to as being “coupled to” or “connected to” another constituent element, one constituent element can be directly coupled to or connected to the other constituent element, but intervening elements may also be present. By contrast, when one constituent element is referred to as being “directly coupled to” or “directly connected to” another constituent element, it should be understood that there are no intervening elements.
[0062] In the present application, it will be appreciated that terms “including” and “having” are intended to designate the existence of characteristics, numbers, steps, operations, constituent elements, and components described in the specification or a combination thereof, and do not exclude a possibility of the existence or addition of one or more other characteristics, numbers, steps, operations, constituent elements, and components, or a combination thereof in advance.
[0063]
[0064]
[0065] As illustrated in
[0066] The NM 11 and the NR 13 are immobile nanoparticles that function as a molecular information storage device and an output unit. An instruction molecular set that is software may be composed of a set of instruction DNAs 2. The set of instruction DNAs 2 programs logical operation using a molecular input storage state of the NM 11 and a kinetics difference between nanoparticle reactions. The relative reaction kinetics for the nanoparticles is adjustable according to DNA hybridization domains bonded to the NM, NF, and NR, a nanoparticle concentration, and an instruction DNA concentration. That is, the reaction rate may be optimized by using the DNA hybridization domains, the nanoparticle concentration, and the instruction DNA concentration.
[0067] Like “Programming using competitive reaction” illustrated at the right-most side of
[0068] The software may be composed of a series of instruction DNAs 2 in a solution. For example, two types of instruction DNAs, the trap DNA, and the report DNA control the NF 12 to bind to the NM 11 or the NR 13 through the DNA hybridization. The instruction DNA 2 controls the NF 12 to bind to the NM 11 or the NR 13 with a different rate according to a different storage state. Accordingly, a mixture of the plurality of instruction DNAs 2 may generate a logical decision-making strategy according to a storage condition. Then, arbitrary logical computing may be programmed in the single LNT hardware chip 1.
[0069] The solution inside the supported lipid bilayer chamber illustrated at the center of
[0070]
[0071] The operation executed by the NVNA follows three steps below.
[0072] As illustrated in A of
[0073] As illustrated in B of
[0074] In operation “input addition” illustrated in A of
[0075] As illustrated in C of
[0076] Next, as illustrated in A of
[0077] In a “Program addition” operation illustrated in A of
[0078] For example, as illustrated in D of
[0079] As described above, the logic operation may be executed by supplying the instruction signal composed of a combination of instruction DNAs. That is, the logic operation may be performed by providing the instruction signal in the form of the combination of instruction DNAs that initiate the nanoparticle-nanoparticle assembly having the different kinetics according to the NM state. When the output signal is 0, all of the NFs are trapped by the memory device NM, but when a result of the logic operation is 1, the NF binds to the output device through the NF-NR assembly to report the output “1”.
[0080] Subsequently, a reset solution is added to the existing solution to dehybridize all of the non-covalent DNA-DNA base pairings, and the LNT 3 returns to the initial state for the next execution.
[0081] In the “Reset solution addition” operation illustrated in A of
[0082] Hereinafter, a software programming strategy using the instruction DNA will be described.
[0083]
[0084] In
[0085] A curve indicated with a black circle in the binding ratio/time graph of
[0086] Referring to the curve indicated with the black circle in the binding ratio/time graph in
[0087] In the dark-field microscope imaging positioned at the centers of
[0088] For example,
[0089]
[0090] As illustrated in
[0091] The NOT gate function may be expressed with an If-Then-Else program script, and may be directly converted to the molecular instruction DNA code composed of the trap DNAs and the report DNAs.
[0092] For example, (if) if the input A is 1, (Then) the NF is trapped to the NM, and (Else) if the input A is not 1, the NF is assembled to the NR. When molecular compiling is performed, an M.sub.1 trap DNA 61 and a report DNA 62 that is the instruction DNAs are determined, and a combination of instruction DNAs that code the nanoparticle neural network is determined through weight coding.
[0093] As illustrated at the rightmost side of
[0094]
[0095] As illustrated in A of
[0096] As monitored with the dark-field microscope, when the input A is 0, two NRs with the scattering color of green are indicated with the dotted circle at 0 minute. It can be seen that when the combination of instruction DNAs programmed with “NOT A” is added, the reporting occurs more than trapping, so that the NR and the NF are assembled (NR-NF) by the report DNA that is one of the instruction DNAs and the scattering color of green-blue is observed after 10 minutes, and the occurrence of the NR-NF assembly increases after 30 minutes. Then, a reporting rate monitored with the scattering color of green-blue by the dark-field microscope is larger than a predetermined threshold of 0.2, so that the output is determined to 1.
[0097] When the input A is 1, one NR with the scattering color of blue and two green NM.sub.A1 are indicated by a dotted circle at 0 minute. In the case where the input A is 1, when the combination of instruction DNAs programmed with “NOT A” is added, the trapping occurs more than the reporting (Trapping>Reporting), the NM.sub.A1 binds to the NF (MA1-NF) by the M.sub.1 trap DNA that is one of the instruction DNAs, so that two scattering color of yellow are observed after 10 minutes. Then, the reporting rate monitored with the scattering color of yellow by the dark-field microscope almost reaches 1, and the reporting rate monitored with the scattering color of green-blue is close to 0 that is smaller than the threshold of 0.2, so that the output is determined to 0.
[0098]
[0099]
[0100]
[0101] In the NVNA-LNT, the program, that is, the set of instructions for the logic circuit, may be composed of a combination of the plurality of trap DNAs and report DNAs.
[0102] In order to enable the nanoparticle to make a logical decision, the exemplary embodiment of the present invention uses two types of NM.sub.0 trap DNA 41 and the NM.sub.1 trap DNA 42 in order to bind the NF to each of the NM.sub.0 and the NM.sub.1. However, the invention is not limited thereto.
[0103] The hybridization domains in the Trap DNA (a.sub.0s, a.sub.1, and f.sub.1) are 14 nucleotides whose melting temperatures are higher than 40° C., allowing fast trapping. On the other hand, another type of instruction DNA, namely the report DNA, is designed to make the NFs be assembled to the NR much more slowly than the logic-allowed trapping for difference in binding kinetics. As illustrated in
[0104] For the report DNA, in the exemplary embodiment, domain r*, which has an 8-base sequence and has a lower melting temperature of about 16° C., is used. In this case, as illustrated in
[0105] For all nanoparticle reactions, detailed experimental conditions (reaction time, DNA concentrations, and buffer conditions), DNA design (domain length, sequence), and kinetics data (τ ½, k.sub.i) will be described below.
[0106] In
[0107] As illustrated in
[0108] As illustrated in
[0109] Further, the M.sub.1 trap DNA shows faster kinetics than the M.sub.0 trap DNA, which is because the a1 domain strands having the low DNA density (low electrostatic repulsion) are exposed over the outer spherical shell of the DNA on the nanoparticle NM M.sub.1.
[0110] Like the upper drawing of
[0111] Referring back to
[0112] When a mixture of the trap DNA and the report DNA is introduced into the NVNA-LNT chip, the NF binding generates a different result according to the NM storage state and the logical operation and implement the logic operation. In case of the output “0”, the NM particle serves as the kinetic trap consuming all of the NFs, so that the NF cannot bind to the NR, resulting in outputting the output “0”. When the result of the logic operation for the NM in which the input is stored is TRUE, the NF is assembled to the NR, and a distinct Plasmon bond is created between the green scattering gold nanoparticle NF and the glue scattering silver nanoparticle NR and the output “1” is output.
[0113] As illustrated in
[0114] The If-Then statement of the corresponding script is substituted to the function of the M.sub.1 trap DNA 61 identically to the method illustrated in
[0115] As illustrated in
[0116]
[0117] After the input DNA is stored in the NM, the chamber was washed with 1×PBS, and cultured for 12 hours at 25° C. In
[0118] As illustrated in
[0119] As described above, in the exemplary embodiment, the data storage operation and the information processing operation may be separated, so that it is possible to easily provide the logical function (for example, the NOT gate function). According to the exemplary embodiment, a predetermined logical gate may be executed without dual-rail transition logic disclosed in the lipid nanotablet (LNT) of Korean Patent Application No. 10-2018-0061345, so that the computing performance of the system may be improved without property reversal.
[0120] Together with the NM.sub.0 trap DNA, the logically inhibited assembly may occur between the NF and the NM.sub.1. Although the input DNA was designed to be trapped in the a.sub.0 domain (24 nucleotides, see
[0121] The reaction network between the plurality of nanoparticles connected by the instruction DNAs may be expressed as a perceptron that is a type of artificial neural network for a binary classifier. Accordingly, in the exemplary embodiment, the programming strategy for configuring the nanoparticle neural network NNN in the LNT platform may be expanded. In this case, the NM, the NF, and the NR may correspond to the input layer, the hidden layer, and the output layer, and the instruction DNA may correspond to the weight concept. The method of determining, by each floater, true or false according to the memory state is similar to the activation function of activating the hidden layer only when the sum of the cross product of the input and the weight (hereinafter, the weight input sum) exceeds a specific threshold. The quantitative value of the weight for the NNN is illustrated in
[0122]
[0123] As illustrated in
[0124] The input layer includes three input nodes M1, M2, and M3, the hidden layer includes four hidden nodes F1, F2, F3, and F4, and the output layer includes one output node R. The NF calculates the weighted sum of inputs in which the weight is multiplied to the inputs and bias upon input condition and is activated with the activation function of the Heaviside step function. The NM.sub.0 trap DNA inactivates the NF in the input “0” and the NM.sub.1 trap DNA inactivates the NF in the input “1”, so that the NM.sub.0 trap DNA and the NM.sub.1 trap DNA may be expressed with discrete weights of 1 and −1, respectively. In order to set the threshold for the activation function to 0, a bias is required to balance the positive and negative values of the sum of weighted inputs. The bias may be defined with the number of NM.sub.0 trap DNAs. The activated NF may bind to the NR with the output “1”.
[0125] The input x.sub.i is 0 or 1, and the weight value w.sub.i,j is 1, 0, or −1. The bias value bj is the number of w.sub.i,j having the weight of 1, and when the sum of the weight input sum (Σ.sub.i=1.sup.nWij.Math.X.sub.i) and the bias vale bj is equal to or larger than 0, the weight v.sub.j between the hidden layer and the output layer is 1, and when the sum of the weight input sum (Σ.sub.i=1.sup.nWij.Math.X.sub.i) and the bias vale bj is smaller than 0, the weight v.sub.j between the hidden layer and the output layer is 0. When the sum of weights v.sub.j is larger than 0, the output y is 1, and when the sum of weights v.sub.j is not larger than 0, the output y is 0.
[0126] Hereinafter, the 2-input Boolean logic gate programming and reset operation by using the nanoparticle neural network will be described.
[0127]
[0128]
[0129]
[0130] There are four input combinations 00, 01, 10, and 11 according to two inputs A and B, and each logic gate determines an output according to the input, and the outputs of the logic gates implemented with the single-layer perceptron of
[0131] In
[0132] In
[0133] The AND logic gate illustrated in
[0134] In
[0135] The INHIBIT logic gate (hereinafter, the INH logic gate) illustrated in
[0136] In
[0137] The NOR logic gate illustrated in
[0138]
[0139]
[0140]
[0141]
[0142] There are four input combinations 00, 01, 10, and 11 according to two inputs A and B, and each logic gate determines the output according to the input, and
[0143] In
[0144] In
[0145] The OR gate illustrated in
[0146] The reporting ratio graphs 181 to 184 for the four input combinations 00, 01, 10, and 11 according to two inputs are illustrated in
[0147] In
[0148] The NAND gate illustrated in
[0149] The reporting ratio graphs 191 to 194 for the four input combinations 00, 01, 10, and 11 according to two inputs are illustrated in
[0150] In
[0151] The XOR gate illustrated in
[0152] The reporting ratio graphs 201 to 204 for the four input combinations 00, 01, 10, and 11 according to two inputs are illustrated in
[0153] In
[0154] The XNOR gate illustrated in
[0155] The reporting ratio graphs 211 to 214 for the four input combinations 00, 01, 10, and 11 according to two inputs are illustrated in
[0156]
[0157] Hardware relaying on the covalently bonded nanostructure on a lipid chip allows for multiple executions by returning back to the initial state by the reset function for reuse. The hatched section illustrated in
[0158] The melting temperature of the DNA domain is considerably lower than the melting temperature under the reset condition, so that the inputs stored in almost all of the instruction DNAs and NMs added to the logic operation may be detached at the low salt concentration and high temperature as described above. After the reset, only the thiolated DNAs remain on the nanoparticles, and the LNT returns to the initial state for the data storage and the logic operation.
[0159] As can be seen in the reporting ratio graph illustrated in
[0160] As described above, it is possible to implement arbitrary Boolean logic circuits in a scalable and modular manner by using the NNN. The software programming function represented in Table 1 below may be provided by executing the functionally complete Boolean logic set for two-bit inputs. Table 1 represents an output of each logic operation according to the trap DNAs TD1 to TD8 and the report DNA RD. In Table 1, “1” represents the insertion of the corresponding trap DNA to the LNG chip, and “0” represents that the corresponding trap DNA is not inserted into the LNT chip. “0/1” represents that the output is determined regardless of the insertion of the DNA RD.
TABLE-US-00001 TABLE 1 NF1 NF1 NF1 NF1 NF2 NF2 NF2 NF2 Output NM.sub.A0 NM.sub.A1 NM.sub.B0 NM.sub.B1 NM.sub.A0 NM.sub.A1 NM.sub.B0 NM.sub.B1 (Input) Trap Trap Trap Trap Trap Trap Trap Trap Report 0, and 0 0, DNA DNA DNA DNA DNA DNA DNA DNA DNA Logic and 1 1, and TD1 TD2 TD3 TD4 TD5 TD6 TD7 TD8 (RD) operation 0 1, and 1 1 1 1 1 1 1 1 1 0/1 False 0 0 0 0 0 0 0 0 0 0 0 0 1 True 1 1 1 1 1 0 0 0 1 1 1 1 1 YES A 0 0 1 1 0 1 0 0 1 1 1 1 1 NOT A 1 1 0 0 0 0 1 0 1 1 1 1 1 YES B 0 1 0 1 0 0 0 1 1 1 1 1 1 NOT B 1 0 1 0 1 0 1 0 1 1 1 1 1 A AND B 0 0 0 1 1 0 0 1 1 1 1 1 1 A AND −B 0 0 1 0 0 1 1 0 1 1 1 1 1 −A AND B 0 1 0 0 0 1 0 1 1 1 1 1 1 A NOR B 1 0 0 0 1 0 0 0 0 1 1 0 1 A OR B 0 1 1 1 1 0 0 0 1 0 0 1 1 A OR −B 1 0 1 1 0 1 0 0 1 0 1 0 1 −A OR B 1 1 0 1 0 1 0 0 1 0 0 1 1 A NAND B 1 1 1 0 1 0 0 1 0 1 1 0 1 A XOR B 0 1 1 0 0 1 0 1 1 1 0 0 1 A XNOR B 1 0 0 1
[0161] The logic gate that yields the single output “1” only in one of the four input combinations, such as AND, INH, and NOR, may be implemented by using the single layer perceptron. As there are two types of NMs, which store inputs A and B, respectively, the NF clearly differentiates the relative kinetics on three binding channels for two NMs and one NR with the input combinations, thereby generating the result of logic gates.
[0162] Next, the multilayer perceptron may be designed by using the plurality of nodes in the hidden layer by introducing the multiple NFs. The plurality of NFs has different DNA domains to trap the trap DNAs, but may share the same domain for binding with the Report DNAs.
[0163] Using the nanoparticle-based multilayer perceptron, the two-input OR, NAND, XOR, and XNOR may be implemented by changing the combination of the trap DNAs. Owing to the two types of NF, the output “1” may have a maximum assembly ratio of approximately 50%. Although the multilayer perceptron exhibited certain density differences between the two NFs, a reporting ratio from 0.2 to 0.6 may represent the output “1,” which indicates a significant difference between “TRUE” and “FALSE”.
[0164] The number of nanoparticle nodes required for functional completeness of the Boolean logic operators may be calculated by the method described below.
[0165] For example, logic circuits dealing with n-bit inputs have 2.sup.n input combinations, each of which may have an output of TRUE or FALSE. Thus, the possible number of logic circuits is 2.sup.2.sup.
[0166] The concept for executing the neural network is proved and the reusability of the NVNA is confirmed through the foregoing description.
[0167] The NNN system according to the exemplary embodiment is combined with the reset function, so that the system may be additionally utilized through a sequential decision-making process using a decision tree.
[0168]
[0169] As illustrated in
[0170] As illustrated in D of
[0171] First, the first decision-making node, “Is input AB greater than input BC?” represented with the electric logic circuit diagram may be the NNN diagram with four nodes in the hidden layer.
[0172] (A) of
[0173] (B) of
[0174] The results of the 16 input cases illustrated in (C) of
[0175] As illustrated in (D) of
[0176] As illustrated in (E) of
[0177] The input of Chip #1 of (E) of
[0178]
[0179] As illustrated in
[0180] The input of Chip #2 of (E) of
[0181] The input of Chip #3 of (E) of
[0182] In the LNT, nanoscale geometric features and optical properties of the plasmonic nanoparticle core are critical for computing. The time-delayed assembly between the NF and NR with the report DNA is attributable to the DNA sequence design and multivalent DNA strand interactions on the single nanoparticle surface. Furthermore, the number of outputs may be increased by adding other pairs of nanoparticle interactions with different scattering colors. With the introduction of blue-scattering NFs along with green-scattering NFs, fan-out logic circuits that operate YES and NOT gates simultaneously with two outputs distinguished by optical signals may be executed.
[0183]
[0184] (A) of
[0185] (B) of
[0186] (C) of
[0187] The B-NF shows that the reaction rate is slow in the reporting. This is due to the low DNA density of the silver nanoparticles. The B-NF is reported slowly, but generates the output “1” or “0” according to the input and the operation of the NOT gate.
[0188] Further, the various characteristics and functions of the nanoparticles and the ligands may be utilized in the computing system. For example, the photothermal characteristic may implement the light-induced elimination function through dehybridization of the loaded molecule, or the magnetic characteristic may enhance the reaction rate by controlling the local concentration of NF spatiotemporally.
[0189] In the exemplary embodiment, it is possible to simplify the basic reaction modules by only using the hybridization between single stranded DNAs through kinetics control, without relying on enzymatic or strand displacement reaction. Therefore, the computing strategy in LNT can be intuitively applicable to other types of molecules, such as proteins, peptides, ionic species, small chemicals, and RNA. As the main purpose of the instruction DNA is to create and control links between nanoparticles with molecular information, other chimeric linkers, such as DNA-antibody or chemical ligand-aptamer conjugates, may serve as the Instruction molecules with binding moieties on both ends.
[0190] The LNT platform according to the exemplary embodiment may be expanded in various directions following the development of neural networks. In the aspect of NNN, instead of using digital values of inputs and weights, varying the weights and inputs to analogue values can be realized with the biochemical molecules via controlling kinetics and ligand density/modification, enabling sophisticated neural network operations for molecular pattern recognition. The number of NFs required for functional completeness exponentially increases as the number of inputs increases, but it is possible to solve the issue by adopting more than one hidden layer by which the fewer number of NFs can cover the same function.
[0191] So far, it has been described that the NVNA implemented in the LNT platform according to the exemplary embodiment can be widely applied and provides the scalable molecular computing. It is possible to realize the NNN having the simple reset function through the computing architecture according to the exemplary embodiment. When the NVNA is constructed, the nanoparticle memory, the instruction DNA, and the DNA binding rate-control output are used, and the solution serves to store, process, and reset information similar to the bus concept of electronic computing.
[0192] The NNN composed of the perceptron, the hidden layer, and the like according to the exemplary embodiment is realized to the NVNA of the lipid chip. The NVNA system may perform the sequential decision-making process according to the decision-making tree, and may be combined with the reset function and reused.
[0193] In the present disclosure, the nanoparticle-based computing architecture is established, and the NNN provides modulated and scalable molecular computing with the single nanoparticle set and the sub set of the instruction DNAs. In general, the nano devices uses new nanoparticle designs for specific functions as their structures define their functions in nanotechnology/life science. However, the exemplary embodiment provides complete diversity of functions together with the nanoparticle network. When the NVNA is used, powerful, scalable, and practical programming functions may be provided in a wide range nanoparticle-based computing application programs. The NNN may provide opportunities to use nanoparticles in deep learning-based science and technology, neural interfaces, and neural morphology computing.
[0194] The exemplary embodiment may provide various intelligent molecular nano computing systems, such as intelligent diagnostic nano devices capable of managing and analyzing complex biomolecular information. For example, consecutive decisions may be made from blood samples for cancer diagnosis, such as “Is it cancer?”, “Is it lung cancer?”, “is lung cancer subtype is 1”, and “How much has lung cancer subtype 1 progressed?”. Further, the LNT chip may be integrated with the compact NVNA computing architecture of a microfluidic device, and the intricate network of interactions among molecules in the solution and nanoparticles on the lipid chip may be used to mimic and interrogate complex living systems.
[0195] The domain and base sequence of the strand bound to the surface of each of the NMs, that is, M.sub.A, M.sub.B, Mc, and M.sub.D, the NFs, that is F.sub.1, F.sub.2, F.sub.3, and F.sub.4, and the NR, that is, R, illustrated in
[0196] The corresponding strands for binding to the corresponding NM are bound to 59% of the surface of each of the NF.sub.1 to NF.sub.4, the same NF-Reporting strands for reporting are bound to 40% of the surface of each of the NF.sub.1 to NF.sub.4, and the NF biotin strands to be attached to the supported lipid bilayer are bound 1% of the surface of each of the NF.sub.1 to NF.sub.4.
[0197] The corresponding strands for binding to the corresponding NF are bound to 60% of the surface of each of the NM.sub.1 to NM.sub.4, and the NM biotin strands to be attached to the supported lipid bilayer are bound to 40% of the surface of each of the NM.sub.1 to NW.
[0198] The strands for binding to the NF are bound to 60% of the surface of the NR, and the NM biotin strands to be attached to the supported lipid bilayer are bound to 40% of the surface of the NR.
TABLE-US-00002 (TABLE 2) Strand (ratio domain Sequence on NPs) (nt) (5' to 3') NF1 59% SH-A 15- Thiol-AAAAAA dummy-f1 14 AAAAAAAAAAAC TGCCATTCATTT ACCGTCTAC (Thiol- SEQ ID NO: 1) NF2 59% SH-A 15- Thiol-AAAAAA dummy-f2 14 AAAAAAAAATAA ACCCATGCTAAC CGCTAACAC (Thiol- SEQ ID NO: 2) NF3 59% SH-A 15- Thiol-AAAAAA dummy-f3 14 AAAAAAAAATCT TAAAGTCCCGTC ACAGTCTAC (Thiol- SEQ ID NO: 3) NF4 59% SH-A 15- Thiol-AAAAAA dummy-f4 14 AAAAAAAAAACC AAAGACCCAAGA TACTCGCAC (Thiol- SEQ ID NO: 4) NF- SH-A 15- Thiol-AAAAAAA Reporting dummy-fr 14 AAAAAAAACGCAC 40% TAAGACGAATTCA CTTTTC (Thiol- SEQ ID NO: 5) NF SH-A 15- Thiol-AAAAAAA biotin dummy- AAAAAAAAATAAC 1% biotin GAGCATCGAACTT ACATCGCGTTACT ACACTGTCACTGA TCATCGCATGCTA TAC-biotin (Thiol- SEQ ID NO: 6- biotin) NM.sub.A 60% a0 24-A CACAACATAGAAT 15-SH CACGTACATAAAA AAAAAAAAAAAA A-Thiol (SEQ ID NO: 7- Thiol) NM.sub.B 60% b0 24-A CCGTTCTACAAAT 15-SH CAAAGCGCATCAA AAAAAAAAAAAA A-Thiol (SEQ ID NO: 8- Thiol) NM.sub.C 60% c0 24-A CTCGTCACTCTAG 15-SH CACAATCTCGAAA AAAAAAAAAAAA A-Thiol (SEQ ID NO: 9- Thiol) NM.sub.D 60% d0 24-A CATAATCTATAAT 15-SH CATCCTCATAAAA AAAAAAAAAAAA A-Thiol (SEQ ID NO: 10- Thiol) NM biotin- biotin-CTCTCT biotin dummy- GCCTCGTTCAGAC 40% A 15-SH AAAACTCATCCTA CTAAAAAAAAAAA AAAA-Thiol (biotin- SEQ ID NO: 11- Thiol) NR r8-dummy- CAATAACGACAAA 60% A 15-SH CTAATTCAGCAAA AAAAAAAAAAAA A-Thiol (SEQ ID NO: 12-Thiol) NR biotin- biotin-CTCTCT biotin dummy-A GCCTCGTTCAGAC 40% 15-SH AAAACTCATCCTA CTCGATACTTTCT GTAACTTTTCCTT AAAAAAAAAAAAA AAAA-Thiol (biotin- SEQ ID NO: 13-Thiol)
[0199] Further, the domain and the base sequence of each of the inputs A, B, C, and D illustrated in
TABLE-US-00003 (TABLE 3) SEQ domain Sequence ID Strand (nt) (5′ to 3′) NO: Input A a0* 24-a1* 14 TTATGTACGTGA 14 TTCTATGTTGTG GATTGTTTTCTA GG Input B b0* 24-b1* 14 GATGCGCTTTGAT 15 TTGTAGAACGGGA TAGGCGACATAG Input C c0* 24-c1* 14 TCGAGATTGTGCT 16 AGAGTGACGAGGA TGTCAGGGACAG Input D d0* 24-d1* 14 TTATGAGGATGAT 17 TATAGATTATGGT ATTTAGAAAATG
[0200] Further, the domain and the base sequence of each of the instruction DNAs illustrated in
TABLE-US-00004 (TABLE 4) SEQ Sequence ID Strand domain (nt) (5' to 3') NO: Report r* 8-dummy- CGTTATTGAA 18 DNA fr* 14 TCCTTAGAGA AAAGTGAATT CG NF1 NM.sub.A0 a0s* 14-dummy- GATTCTATGT 19 Trap DNA f1* 14 TGTGCCTTCA TTAAGTAGAC GGTAAATG NF1 a1 14-dummy- CCTAGAAAAC 20 NM.sub.A1Trap f1* 14 AATCTCTCGT DNA ATTCGTAGAC GGTAAATG NF1 NM.sub.B0 b0s* 14-dummy- GATTTGTAGA 21 Trap DNA f1* 14 ACGGATATCA TTACGTAGAC GGTAAATG NF1 NM.sub.B1 b1 14-dummy- CTATGTCGCC 22 Trap DNA f1* 14 TATCCATATA CAGCGTAGAC GGTAAATG NF1 NM.sub.C0 c0s* 14-dummy- GCTAGAGTGA 23 Trap DNA f1* 14 CGAGAAGATA TACCGTAGAC GGTAAATG NF1 NM.sub.C1 c1 14-dummy- CTGTCCCTGA 24 Trap DNA f1* 14 CATCTTCTTA TCTCGTAGAC GGTAAATG NF1 NM.sub.D0 d0* 14-dummy- GATTATAGAT 25 Trap DNA f1* 14 TATGCCTACG TCTAGTAGAC GGTAAATG NF1 NM.sub.D1 d1 14-dummy- CATTTTCTAA 26 Trap DNA f1* 14 ATACACAGCT AACAGTAGAC GGTAAATG NF2 NM.sub.A0 a0s* 14-dummy- GATTCTATGT 27 Trap DNA f2* 14 TGTGTTCTTT ACCAGTGTTA GCGGTTAG NF2 a1 14-dummy-f2* CCTAGAAAAC 28 NM.sub.A1Trap 14 AATCTATGCC DNA TACTGTGTTA GCGGTTAG NF2 NM.sub.B0 b0s* 14-dummy- GATTTGTAGA 29 Trap DNA f2* 14 ACGGAACATT ACCAGTGTTA GCGGTTAG NF2 NMb1 b1 14-dummy- CTATGTCGCC 30 Trap DNA f2* 14 TATCTTAAAC CGTAGTGTTA GCGGTTAG NF2 NM.sub.C0 c0s* 14-dummy- GCTAGAGTGA 31 Trap DNA f2* 14 CGAGATACCA AATAGTGTTA GCGGTTAG NF2 NM.sub.C1 c1 14-dummy- CTGTCCCTGA 32 Trap DNA f2* 14 CATCACTACT ATTTGTGTTA GCGGTTAG NF2 NM.sub.D0 d0* 14-dummy- GATTATAGAT 33 Trap DNA f2* 14 TATGACAGTC GCAGGTGTTA GCGGTTAG NF2 NM.sub.D1 d1 14-dummy- CATTTTCTAA 34 Trap DNA f2* 14 ATACCACGGC CTTGGTGTTA GCGGTTAG NF3 NM.sub.A0 a0s* 14-dummy- GATTCTATGT 35 Trap DNA f3* 14 TGTGAGCTTC ATGCGTAGAC TGTGACGG NF3 NM.sub.A1 a1 14-dummy- CCTAGAAAAC 36 Trap DNA f3* 14 AATCTCCAGT CTTTGTAGAC TGTGACGG NF3 NM.sub.B0 b0s* 14-dummy- GATTTGTAGA 37 Trap DNA f3* 14 ACGGAACCTA GTAAGTAGAC TGTGACGG NF3 NM.sub.B1 b1 14-dummy- CTATGTCGCC 38 Trap DNA f3* 14 TATCTAACCT TACAGTAGAC TGTGACGG NF3 NM.sub.C0 c0s* 14-dummy- GCTAGAGTGA 39 Trap DNA f3* 14 CGAGCTTGCA AACAGTAGAC TGTGACGG NF3 NM.sub.C1 c1 14-dummy- CTGTCCCTGA 40 Trap DNA f3* 14 CATCTAATCA AGCAGTAGAC TGTGACGG NF3 NM.sub.D0 d0s* 14-dummy- GATTATAGAT 41 Trap DNA f3* 14 TATGAACCTA GTAA NF3 NM.sub.D1 d1 14-dummy-f3* CATTTTCTAA 42 Trap DNA 14 ATACTAACCT TACAGTAGAC TGTGACGG NF4 NM.sub.A0 a0s* 14-dummy- GATTCTATGT 43 Trap DNA f4* 14 TGTGTTACAA TAACGTGCGA GTATCTTG NF4 a1 14-dummy-f4* CCTAGAAAAC 44 NM.sub.A1Trap 14 AATCTATCAC DNA ACAGGTGCGA GTATCTTG NF4 NM.sub.B0 b0s* 14-dummy- GATTTGTAGA 45 Trap DNA f4* 14 ACGGACACTT TGCGGTGCGA GTATCTTG NF4 NM.sub.B1 b1 14-dummy-f4* CTATGTCGCC 46 Trap DNA 14 TATCCAAATA GAATGTGCGA GTATCTTG NF4 NM.sub.C0 c0s* 14-dummy- GCTAGAGTGA 47 Trap DNA f4* 14 CGAGATATCA CAAGGTGCGA GTATCTTG NF4 NM.sub.C1 c1 14-dummy-f4* CTGTCCCTGA 48 Trap DNA 14 CATCTACTTA CGATGTGCGA GTATCTTG NF4 NM.sub.D0 d0s* 14-dummy- GATTATAGAT 49 Trap DNA f4* 14 TATGACACTT TGCGGTGCGA GTATCTTG NF4 NM.sub.D1 d1 14-dummy-f4* CATTTTCTA 50 Trap DNA 14 AATACCAAA TAGAATGTG CGAGTATCT TG
[0201] The Trap DNAs includes trap DNAs which binds each of the NF.sub.1 to NF.sub.4 to one of the NM.sub.A to NM.sub.D in which the storage state is 0 or 1. The report DNA binds each of the NF.sub.1 to NF.sub.4 to the NR.
Preparation of Small Unilamellar Vesicles (SUVs)
[0202] After sonication cleaning of a 50 mL round bottom flask with 99.5% chloroform (DAEJUNG, South Korea), 97.2 mol % dioleoylphosphatidylcholine (DOPC), 0.4 mol % biotinylated dioleoylphosphatidylethanolamine (DOPE), and 2.5 mol % poly(ethylene glycol) 1000 (PEG 1000)-DOPE (all three lipids were purchased from Avanti, USA) were mixed in the solution of 99.8% chloroform (SAMCHUN, South Korea). Chloroform solvent was removed via a rotary evaporator, which left a lipid mixture ring film in the round bottom flask. To ensure removal of all chloroform, N.sub.2 blowing for 5 min was followed. The mixture was re-suspended in de-ionized water (DIW), resulting in lipid solution in the concentration of 2 mg/mL. The solution was taken three freeze-thaw steps from −76° C. to 25° C. and kept in liquid nitrogen. To make small unilamellar vesicles (SUVs), after unfreeze the lipid solution at 25° C., we conducted 30-min sonication before use.
Preparation of Supported Lipid Bilayer (SLB) Chamber
[0203] SLB was formed on hydrophilic supporting substrates. We used the vesicle fusion method to get SLBs in the flow chamber. We used flow chamber consisting of top slide glass, Parafilm spacer (4 mm×50 mm×200 μm), and bottom cover glass (both glasses are purchased from Paul Marienfeld GmbH & Co. KG, Germany). The inner volume of the flow chamber is ˜40 μL. The top slide glass, which has inlet and outlet holes, and the bottom cover glass were cleaned by 10 min sonication in DIW and 10 min piranha etching in H.sub.2SO.sub.4/H.sub.2O.sub.2 (3:1), and were thoroughly rinsed with DIW. To prevent SLB formation on the top glass, we passivated the top slide glass with bovine serum albumin (BSA) (10 mg/mL) in 150 mM NaCl phosphate-buffered saline (1×PBS) for 30 min. The flow chamber was assembled with placing double layer Parafilm spacer between the two glasses and heat sealing at 105° C. Next, the SUV solution was diluted to 1 mg/mL in 1×PBS solution and sonicated additionally for 15 min. The SUV solution is injected into the flow chamber and incubated for 40 min to form a lipid bilayer on the bottom of the chamber. The flow chamber was gently washed by injection of DIW (200 μL, twice) and 1×PBS (200 μL, once), and passivated with BSA (30 μg/mL) in 1×PBS for 30 min. Streptavidin (40 nM) in 1×PBS was injected into the flow chamber and incubated for 90 min to modify biotinylated DOPE in SLBs. Finally, the flow chamber was washed with 1×PBS (200 μL, twice), and modified with DNA-functionalized nanoparticles for further experiments.
[0204] Synthesis and Characterization of Plasmonic Nanoparticles
[0205] Spherical gold nanoparticles (AuNPs) (diameter, 50.5±3.5 nm) for green-scattering color were purchased from BBI Solutions (Cardiff, UK). Gold-silver core-shell nanoparticles (Au@Ag NPs) (diameter, 50.3±4.7 nm) for blue-scattering color were synthesized by following seed-mediated growth method. We prepared CTAB-capped seeds. Cetyltrimethylammonium bromide (CTAB) solution (9.75 mL, 100 mM) was mixed with HAuCl.sub.4 solution (250 μL, 10 mM). Freshly made ice-cold NaBH.sub.4 solution (600 μL, 10 mM) was added quickly to the mixture with vigorous stirring for 3 min. The synthesized 1-2 nm seeds were incubated at 27° C. for 3 h before the next step. 10-nm gold core nanoparticles were then synthesized with the seeds. CTAC (2 mL, 200 mM), L-ascorbic acid (1.5 mL, 100 mM), and the previously prepared CTAB-capped seed solution (50 μL) were sequentially mixed. HAuCl.sub.4 solution (2 mL, 0.5 mM) was injected with one shot while the solution was being mixed and incubated at 25° C. for 15 min with stirring at 300 rpm. After incubation, the 10-nm gold cores were washed by centrifugation and redispersed in cetyltrimethylammonium chloride (CTAC) solution (1 mL, 20 mM). Finally, we grew a silver shell on the gold core. We mixed the gold core solution (5 μL, 1.77 pM) with CTAC solution (100 μL, 100 mM), AgNO.sub.3 solution (30 μL, 1 mM), and NH.sub.4OH solution (5 μL, ACS reagent, 28.0-30.0% NH.sub.3 basis) sequentially. After gold core solution and L-ascorbic acid solution (100 mM) were mixed and heated to 50° C., the L-ascorbic acid solution (50 μL) was quickly injected and rapidly mixed. The solution was washed via centrifugation and redispersed in 1% polyvinylpyrrolidone (PVP) solution for further oligonucleotide modification. The grown silver shell thickness was ˜20 nm. All NPs were characterized by transmission electron microscopy (TEM) (JEM-2100, JEOL Ltd., Japan), UV-Vis spectrophotometry (Agilent 8453, Agilent Technologies, USA) and DFM (Axiovert 200 M, Carl Zeiss, Göttingen, Germany). TEM imaging was carried out at the National Center for Inter-University Research Facilities (Seoul National University, Seoul, South Korea).
[0206] Functionalization of Plasmonic Nanoparticles
[0207] Gold and silver nanoparticles are modified with thiolated DNA oligonucleotides via strong gold-thiol and silver-thiol bond. To cleave the dithiol bond, thiol modified oligonucleotides (Integrated DNA Technologies, USA) are incubated with dithiothreitol (100 mM) in pH 8.0 phosphate buffer for 1 hour. The DNA was purified through size exclusion chromatography using a NAP-5 column (GE Healthcare, Buckinghamshire, UK), and the concentration of monothiolated DNA was measured by UV-Vis spectroscopy. Nanoparticles (10 fmole) were mixed with the thiolated oligonucleotides (200 pmole), at 0.1% (w/v) sodium dodecyl sulfate (SDS) solution (300 mL), and incubated for 1 h at 25° C. The ratios of thiolated strands for Nano-Floater (NF), Nano-Memory (NM), Nano-Reporter (NR) are summarized in Table S2. Three aliquots of 1 M NaCl, 0.1% SDS, and 10 mM PB salt solution were added with a 1-hour interval to achieve a final concentration of 0.3 M NaCl for NF and NR, and 0.25 M NaCl for NM. The solution was sonicated for 10 seconds after each salt aging, and incubated overnight at 25° C. The nanoparticles were centrifuge-washed and re-dispersed in 10 mM PB solution.
Data Storage on NM and Nanoparticle Neural Network (NNN) Execution on LNT
[0208] NF, NM.sub.1 and NR of which concentration range between 1 to 10 pM in 1×PBS were loaded on LNT for 10 min to reach a proper density of particles (0.1 to 0.2 mm.sup.−2 for NM, ˜0.02 mm.sup.−2 for NF, and ˜0.02 mm.sup.−2 for NR) on SLB. To achieve trapping faster than reporting, we loaded the density of NM 5 to 10 times higher than that of NR. To fully store molecular information of Input DNAs on NM.sub.1 we incubated each Input DNA at 50 nM concentration for 30 min and washed with 1×PBS. The mixture of Instruction DNAs (8 nM Trap DNAs and 1 nM Report DNA) in 1×PBS were injected to the chamber. The nanoparticle logic circuit operation was monitored via DFM for 30 minutes. The sequences of Instruction DNAs are summarized in Table S3.
[0209] To reset the LNT and reuse for the next operation, all the hybridized Input DNAs and the Instruction DNAs need to be removed from the nanoparticles. We increased the temperature and decreased the salt concentration of the solution over the melting temperature to make solution temperature is greater than the melting temperature of DNA hybridization. PB solution (5 mM) was injected, and incubate the LNT chamber at 50° C. for 30 minutes to detach the hybridized DNAs. We then washed with PB solution (5 mM, 50° C.). For the next operation, 1×PBS solution was injected to recover the salt concentration and temperature.
Analyzing Dark-Field Time-Lapse Images
[0210] To analyze obtained dark-field time-lapse images, we used the previously developed custom MATLAB code from our lab (8). Images were registered with The StackReg plugin in ImageJ, choosing the area of interest (100×100 mm.sup.2) and correcting the movement. The image sequences were processed by an image analysis algorithm that enables single-particle signal tracking, signal configuration, and classification. Analyzing the NP's movement and RGB profile, it classifies NPs into NRs, NFs and NMs. The nanoparticle assemblies were quantitatively counted by monitoring scattering color change at the position of NR and NM (immobile NPs) upon NF binding.
[0211] While this invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments. On the contrary, it is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims
DESCRIPTION OF SYMBOLS
[0212] 1: Hardware chip [0213] 2: Software [0214] 3: LNT (lipid nanotablet)