A DNA storage data inverted index retrieval method
Patent Information
- Application Number
- CN202611019228.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-09
- Publication Date
- 2026-10-13
Smart Images
Figure CN122907159A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of bioinformatics and nucleic acid chemistry, specifically to a method for retrieving DNA storage data by inverted index, which is based on the CRISPR / dCas9 system and is used to achieve fine-grained, content-driven, non-destructive retrieval of DNA storage data that supports complex logical queries. Background Technology
[0002] Efficiently acquiring and retrieving massive amounts of data is a core issue in information system design in the digital age. As the cornerstone of data collection, web crawlers, through pre-defined addressing ranges and parsing logic, achieve automated, large-scale crawling and structured reorganization of the digital space. This inverted index collection model, centered on "keyword-document" mapping, has laid the foundation for the basic paradigm of modern large-scale information retrieval systems. Meanwhile, with the rapid development of the DNA field, DNA storage, with its ultra-high density characteristics, provides a new scalable medium for handling the data explosion.
[0003] In order to achieve efficient retrieval of information at the molecular level, DNA information retrieval methods developed by drawing on traditional retrieval logic are becoming a current research hotspot.
[0004] Currently, DNA information retrieval methods mainly revolve around polymerase chain reaction (PCR) and physical matrix encapsulation strategies. These strategies typically require prior knowledge and the assignment of a unique address to each data file to establish an "address write-address read" search process. These strategies have demonstrated a certain accessibility in random access to large-scale information. However, the application and development of existing strategies are still constrained by their retrieval accessibility limitations. Specifically: current mainstream PCR-based methods, to achieve high-precision retrieval, require the introduction of redundant address sequences to cooperate with nested PCR operations, which inevitably compresses the effective logical density of the system; simultaneously, excessive molecular crosstalk within primers and between primers and data payloads limits the throughput of multiplexing; furthermore, high-temperature cycle-dependent PCR access may cause file corruption and abundance shifts. Another strategy, utilizing physical matrices to physically encapsulate and isolate databases, spatially eliminates molecular crosstalk; however, the orthogonal index dimension it relies on is limited by the development of physical matrices, and its encoding capacity is significantly lower than primer design. Furthermore, compared to the single-strand level retrieval granularity of PCR, encapsulation / isolation methods typically only provide library-level retrieval granularity and cannot operate on intra-strand data bytes.
[0005] In recent years, the CRISPR / Cas system (Clustered Regularly Interspaced Short Palindromic Repeats / Cas) has brought transformative opportunities to the field of DNA data storage and retrieval due to its programmable sequence-specific recognition capabilities mediated by single-stranded guide RNA (sgRNA). The CRISPR / Cas system has inspired exploration into its application for DNA data access. Specifically, because the CRISPR / Cas system possesses the ability to search in one-dimensional sliding and three-dimensional collisions in a liquid phase, it has the ability to rapidly locate data in large-scale, complex search spaces. Based on this capability, the cleavage-active Cas12a protein can be used for keyword reporting via fluorescent signals, and selective cleavage using the Cas9 protein can achieve nanopore sequencing enrichment.
[0006] However, existing strategies either only detect the presence of keywords without physically extracting the file, or irreversibly damage the target's DNA during the retrieval process. Furthermore, they generally lack the ability to execute complex, multi-keyword logical queries, limiting their practical application with large-scale information.
[0007] In summary, the existing methods have the following shortcomings: 1) Address-based retrieval methods lack the ability to quickly locate data in large-scale complex search spaces, exhibiting a linear time complexity of O(n) when processing massive amounts of samples. 2) Existing methods are mostly limited to coarse-grained retrieval at the library level. In addition, to ensure retrieval specificity, existing retrieval methods also need to introduce a large number of non-coded address sequences, which not only compresses the effective data carrying space but also increases the redundancy of the system. For example, PCR-based methods require the introduction of redundant address sequences to facilitate nested PCR operations, which inevitably compresses the effective data density of the system; excessive molecular crosstalk within primers and between primers and data payloads limits the throughput of multiplexing; the orthogonal index dimension on which encapsulation / isolation methods rely is limited by the development of physical matrices, and its encoding capacity is significantly lower than that of primer design; encapsulation / isolation methods typically only provide library-level retrieval granularity and cannot operate on intra-chain data bytes. 3) Traditional PCR access relies on high-temperature thermal cycling, and retrieval schemes based on cleavage nucleases can cause irreversible damage to the DNA template. This destructive operation prevents data from being repeatedly retrieved, severely limiting the robustness of DNA as a long-term stable storage medium. For example, early CRISPR retrieval schemes often only achieved fluorescence detection without physical enrichment, or irreversibly damaged the DNA template during the retrieval process, and lacked the ability to perform multi-keyword Boolean logic queries.
[0008] Therefore, it is necessary to propose an inverted index retrieval method for DNA storage data to construct an inverted index (keyword-document mapping) similar to Internet retrieval at the molecular level, and to achieve highly specific, parallel and lossless physical extraction, thereby enabling DNA storage to become practical. Summary of the Invention
[0009] The purpose of this invention is to provide a method for retrieving DNA storage data by inverted index, so as to achieve fast and lossless keyword retrieval and physical enrichment.
[0010] To achieve the above objectives, the present invention provides a method for retrieving DNA storage data using an inverted index, comprising: Construct at least one DNA information chain, including a nucleotide sequence obtained by converting at least one keyword and a PAM motif at the end used to distinguish whether the DNA information chain is compliant; Step S1: Construct a DNA crawler unit based on the target keywords. The DNA crawler unit includes a dCas9 protein without endonuclease activity and a guide nucleic acid chain. The PI domain of the dCas9 protein serves as a PAM navigator, and the guide nucleic acid chain serves as the query statement for the keywords. Step S2: During the index search, the DNA information strand is mixed with the DNA crawler unit; Step S3: Use the PAM navigator to determine whether the end of the DNA information chain has a compliant PAM motif. If so, use a query statement to match the nucleotide sequence corresponding to the target keyword to achieve targeted search for the target keyword. Step S4: Using the physical or chemical characteristics of changes in DNA crawler units after binding to nucleotide sequences, the DNA information strand carrying the target keyword is separated.
[0011] The guide nucleic acid strand is a programmable sgRNA, crRNA, or gRNA.
[0012] The dCas9 protein has a REC domain, which is used to clamp a successfully matched nucleotide sequence corresponding to the target keyword and a guide nucleic acid strand.
[0013] In step S0, multiple DNA information strands carrying different keywords coexist in the same oligonucleotide pool; in step S1, the number of DNA crawler units is one; or, the number of DNA crawler units is multiple, each DNA crawler unit has the same type of dCas9 protein and guide nucleic acid strand, and different DNA crawler units have different guide nucleic acid strands.
[0014] In step S4, when the number of DNA crawler units is multiple, the separated DNA information chain with target keywords is a DNA information chain with multiple target keywords.
[0015] The dCas9 protein is one of SpCas9 protein, VQR-Cas9 protein, VRER-Cas9 protein, and SpRY-Cas9 protein.
[0016] The physical characteristics include at least one of molecular weight, spatial conformation, and charge; the chemical characteristics include specific affinity. The DNA information strands carrying the target keywords are separated, specifically by physically separating the DNA information strands carrying the target keywords from the background pool under isothermal conditions through electrophoretic mobility variation analysis or magnetic bead sorting.
[0017] The guide nucleic acid chain uses programmable sgRNA, and the molar ratio of dCas9 protein to sgRNA is 1:1 to 1:10.
[0018] The DNA storage data inverted index retrieval method of this invention combines the programmability of CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) with the non-destructive, logical, and physical extraction capabilities of a computational search engine to achieve fast, lossless keyword retrieval and physical enrichment. It establishes a new paradigm for interactive, content-driven search within molecular memory, propelling DNA data storage from a static archive to a dynamic, searchable information resource. Attached Figure Description
[0019] Figure 1 This is a schematic diagram illustrating the working principle of the DNA storage data inverted index retrieval method of the present invention.
[0020] Figure 2 This is a schematic diagram of the retrieval logic of the DNA storage data inverted index retrieval method of the present invention.
[0021] Figure 3 This is an experimental schematic diagram illustrating the DNA storage data inverted index retrieval method of the present invention during the specific verification of compliant access.
[0022] Figure 4This is an experimental result diagram of the DNA storage data inverted index retrieval method of the present invention during the specific verification of compliant access.
[0023] Figure 5 This is an experimental schematic diagram of the DNA storage data inverted index retrieval method of the present invention for fuzzy retrieval and Boolean logic retrieval of DNA libraries.
[0024] Figure 6 This figure shows the experimental results of the DNA storage data inverted index retrieval method of the present invention in fuzzy retrieval and Boolean logic retrieval of DNA libraries. Detailed Implementation
[0025] The present invention will be further described in detail below with reference to embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto. The pharmaceuticals and reagents used in the embodiments can be purchased from chemical pharmaceutical companies or prepared by methods disclosed in the art.
[0026] like Figure 1 and Figure 2 As shown, the DNA storage data inverted index retrieval method of the present invention specifically includes the following steps: Step S0: Construct at least one DNA information chain, the DNA information chain comprising a nucleotide sequence obtained by converting at least one keyword and a PAM motif at the end for distinguishing whether the DNA information chain is compliant; To initiate the aforementioned chemosynthesis-driven retrieval, this invention constructs a mapping rule set (codebook) between characters and DNA information chain sequences, and embeds keywords as fixed-length oligonucleotide tags into different DNA information chains. In this embodiment, the oligonucleotide tag length is fixed at 22 bp (of which 2 bp is a GG tag). For a specific sgRNA, the PAM end of the oligonucleotide has approximately 9 bp of mismatch tolerance.
[0027] In this invention, multiple DNA information strands carrying different keywords can coexist in the same oligonucleotide pool.
[0028] Step S1: Construct DNA crawler units based on target keywords. The DNA crawler units adopt the CRISPR / dCas9 system and include dCas9 protein without endonuclease activity and guide nucleic acid strands. The DNA information chain includes nucleotide sequences derived from various keywords, and the keywords needed for indexing and searching are called target keywords.
[0029] The PI domain of the dCas9 protein acts as a PAM navigator, used to determine whether the end of the DNA information strand carries a compliant PAM motif. The dCas9 protein is a mutant protein that retains only sequence recognition function and lacks endonuclease activity (see the literature [Jinek, M. et al. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 337, 816–821 (2012)]), thus it is defined as a non-destructive search executor driving dynamic exploration.
[0030] The guide nucleic acid strand is used as a keyword in a query to achieve targeted searching. In this embodiment, the guide nucleic acid strand is a programmable sgRNA. In other embodiments, the sgRNA can be replaced with crRNA or gRNA.
[0031] The REC domain of the dCas9 protein is used to clamp successfully matched nucleotide sequences corresponding to the target keyword and the guide nucleic acid strand. Furthermore, it facilitates rapid dissociation of the complex in the event of nucleotide sequence and guide nucleic acid strand mismatch. Thus, the REC domain and guide nucleic acid strand of the dCas9 protein together act as a content filter, enabling targeted searching for the target keyword.
[0032] The number of DNA crawler units can be one or more. Each DNA crawler unit has the same type of dCas9 protein and guide nucleic acid strand, while different DNA crawler units have different guide nucleic acid strands. They can be expanded and replaced as needed, thereby solving the problem of retrieving massive amounts of data in DNA storage and providing an efficient and lossless technical path for the dynamic retrieval of massive DNA data.
[0033] Therefore, this invention uses the CRISPR / dCas9 system as a DNA crawler unit. The CRISPR / dCas9 system includes the PI domain of the dCas9 protein as a PAM navigator and a guide nucleic acid strand as a content filter. After successful matching, the ternary complex formed by the combination of the CRISPR / dCas9 system and the DNA information strand serves as a physical downloader. By utilizing the molecular weight and charge difference generated after binding, the desired target DNA is separated, thereby achieving high-fidelity download.
[0034] Step S2: During the index search, the DNA information strand is mixed with the DNA crawler unit; Step S3: Use the PAM navigator to determine whether the end of the DNA information chain has a compliant PAM motif (Protospacer Adjacent Motif). If so, use a query statement to match the nucleotide sequence corresponding to the target keyword to achieve targeted search for the target keyword. This allows DNA information strands carrying only compliant PAM motifs to be retrieved and subsequent processes to be executed, while private DNA strands with base mutations in the PAM sequence are directly filtered out. Thus, access control for compliant information is achieved through the identification of PAM motifs.
[0035] This invention leverages the absolute dependence of the dCas9 protein on PAM motifs to functionalize and map specific base sequences at the ends of the DNA information strand into compliance verification instructions (similar to the Robots.txt protocol in web crawlers). Specifically, a DNA crawler unit is added to the DNA sequence to be tested. When the DNA crawler unit slides and scans across the DNA sequence, the PI domain acts as a PAM navigator, anchoring the PAM sequence on the DNA. Only when the target DNA information strand carries a compliant PAM motif does the PAM motif bind to the PI domain of the dCas9 protein. Only then can the mutated dCas9 protein, acting as the retrieval executor, initiate the recognition process, thereby achieving the concealment of private information and access control over compliant information.
[0036] In this embodiment, the dCas9 protein uses the most commonly used SpCas9 protein to recognize 5'-NGG-3' as a PAM motif. However, this invention is also applicable to other types of Cas9 proteins and can be used to recognize other different types of PAM motifs, such as VQR-Cas9 (recognizing 5'-NGA-3'), VRER-Cas9 (recognizing 5'-NGCG-3'), SpRY-Cas9 (recognizing full PAM), etc.
[0037] As mentioned above, the DNA information chain to be tested includes nucleotide sequences derived from various keywords. The keywords required for indexing and searching are called target keywords. In this invention, the dCas9 complex utilizes a search mechanism combining one-dimensional sliding and three-dimensional collision in the liquid phase to achieve parallel targeted searching of massive DNA libraries. Once the guide nucleic acid chain matches the target keyword, the target sequence at the 5′ end of the programmable sgRNA matches the nucleotide sequence corresponding to the target keyword in the DNA, thus triggering a binding event. Figure 2 ).
[0038] The REC domain of the dCas9 protein is used to clamp successfully matched nucleotide sequences corresponding to the target keyword and the guide nucleic acid strand. Thus, the REC domain and guide nucleic acid strand of the dCas9 protein together act as a content filter, enabling targeted searching for the target keyword.
[0039] Step S4: Utilize the physical or chemical characteristics of changes in DNA crawler units after binding to nucleotide sequences to separate the DNA information strands carrying target keywords, thereby achieving enrichment and downloading.
[0040] This enabled the binding and enrichment download of DNA crawler units with target keywords.
[0041] The physical characteristics include at least one of molecular weight, spatial conformation, and charge, and the chemical characteristics include specific affinity.
[0042] Accordingly, the DNA information strands containing the target keywords are separated, specifically by using electrophoretic mobility shift analysis (EMSA) or magnetic bead sorting to physically separate the DNA information strands containing the target keywords (i.e., the ternary complex formed by the combination of DNA crawler units and DNA information strands) from the background pool under isothermal conditions, thereby achieving enrichment and downloading, and finally restoring them into user-readable text through sequencing decoding.
[0043] Because multiple DNA information strands carrying different keywords coexist in the same oligonucleotide pool in step S0, in step S1, the DNA crawler unit can be one or more types. Each DNA crawler unit has the same type of dCas9 protein and guide nucleic acid strand, but different DNA crawler units have different guide nucleic acid strands. When the query is changed or added, the system can perform a corresponding selective targeted search for the target keyword within the same mixed system. In step S4, when the number of DNA crawler units is multiple, the isolated DNA information strand carrying the target keyword can be a DNA information strand carrying multiple target keywords.
[0044] In this invention, key parameters include: 1) The ratio of dCas9 protein and sgRNA in the CRISPR / dCas9 system: The molar ratio of dCas9 to sgRNA is 1:1 to 1:10. Preferably, the molar ratio of dCas9 to sgRNA is 1:6 (final concentrations of 25 nM and 150 nM, respectively). At this ratio, dCas9 protein can be fully recruited to the navigation site and maintain stable molecular gripping activity.
[0045] 2) The DNA information chain serves as the logical substrate for the DNA crawler unit to grasp and filter. The final concentration of each component of the DNA information chain is usually controlled at 5 nM to balance the reaction state and the gel image performance.
[0046] 3) The reactions in steps S2-S3 are carried out at 25-40 °C, preferably at 37 °C, and in 1×3.1 buffer to maintain the stability of the DNA structure and enzyme activity.
[0047] 4) Electrophoresis detection is usually set to run at 40-180V, preferably at a constant voltage of 60V for 120 minutes, and is combined with an ice bath to improve the resolution of fragments with small molecular weight differences.
[0048] Example 1: DNA Storage Data Inverted Index Retrieval Method for Verifying Compliance Access to PAM Gated Data The DNA storage data inverted index retrieval method proposed in this invention can perform specific information crawling when the DNA information database meets compliance constraints, and will not crawl information when it does not comply.
[0049] The following are the specific experimental results: In step S0, as Figure 3 As shown, two 59 bp DNA information strands were constructed, with their terminal PAM motifs labeled [GG] and [GT], respectively, to correspond to public and private information. The public DNA information strand carries a compliant 5'-GG-3' PAM tag at its end and embeds a 22 bp keyword oligonucleotide tag within the strand; the private DNA information strand has a 5'-GT-3' mutation at its end, and the remaining sequence is completely identical to the public DNA information strand. The public and private DNA information strands were mixed in equimolar amounts to construct a background pool containing a DNA library, with a working concentration of 250 nM for the DNA information strands.
[0050] All the coded experiments and the buffer solutions used for the preparation of the annealing complex contained 12.5 mM Mg. 2+ 1×TE buffer was used. The annealing procedure was the same as that used for other double-stranded complexes. All DNA strands were synthesized by commercial synthetic companies (such as Sangon Biotech) using the solid-phase phosphoramide method. The synthesized products were purified by high-performance liquid chromatography to remove short-chain impurities and byproducts generated during synthesis. The purified DNA strands were identified by electrospray ionization mass spectrometry, confirming that the molecular weight was consistent with the theoretical value and the purity was ≥95%. Each purified DNA information strand was dissolved in 1×TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) to prepare a 100 μM stock solution, which was stored at -20℃ for later use.
[0051] In step S1, when constructing the DNA crawler unit, the total reaction volume is 25 μL, specifically including: 1.25 μL sgRNA (3 μM), 1 μL dCas9 (625 nM), 2.5 μL 10× 3.1 buffer, and 18.75 μL ultrapure water added to a 50 μL EP tube, and thoroughly premixed; then, incubated at 40°C on the PCR instrument top and 37°C on the system for 30 min to allow dCas9 and sgRNA to fully bind and form a binary complex; In step S2, 1.5 μL of dsDNA information strand with a concentration of 250 nM is added to the reaction system containing DNA crawler units.
[0052] In step S3, the mixture is kept at 37 °C for 30 minutes to achieve binding incubation, complete PAM navigator recognition, target keyword matching, and ternary complex (dCas9-sgRNA-DNA) assembly.
[0053] In addition, three parallel experimental groups were set up at 25℃, 30℃, and 40℃, with other parameters remaining unchanged, to verify that specific capture can be achieved at temperatures ranging from 25℃ to 40℃.
[0054] In step S4, the DNA information strand carrying the target keyword is physically separated from the background pool in the form of a dCas9-sgRNA-DNA ternary complex by electrophoretic mobility shift analysis (EMSA).
[0055] The sample preparation method for electrophoresis analysis was as follows: Take 25 μL of the sample obtained in step S3 after the reaction, add 2.5 μL of 10× DNA loading buffer, mix thoroughly by pipetting, and centrifuge. Prepare a 1% agarose gel, mix the nucleic acid dye thoroughly, and cool the electrophoresis tank on ice. For sample loading, add 20 μL of the sample and 2 μL of nucleic acid molecular weight marker to each well. Perform electrophoresis on ice to increase resolution. Electrophoresis parameters: set voltages of 40 V, 80 V, 120 V, and 180 V; the baseline condition was a constant voltage of 60 V for 120 minutes.
[0056] After electrophoresis, the gel was placed under a UV gel imaging system to take gel photographs and read the dsDNA information strand crawling status.
[0057] like Figure 4As shown, we can see the public DNA information strand, the private DNA information strand, and the mixed DNA information strand. Among them, the dCas9-sgRNA-DNA ternary complex is significantly shifted upward due to migration hindrance caused by changes in molecular weight and charge. Only when the [GG] marker is completely preserved at the end of the information strand can the DNA crawler accurately identify the public information and achieve physical separation in electrophoresis. Figure 4 This means that only public information marked with [GG] was crawled; private information chains marked with [GT] were not identified and were therefore excluded. Once a single base change occurs at this site (such as G→T), identification fails, effectively "hiding" the private information in the background pool.
[0058] Example 2: A DNA Storage Data Inverted Index Retrieval Method for Parallel Logical Retrieval of DNA Libraries Parallel logical retrieval refers to the ability to use multiple different sgRNAs in the same system to simultaneously retrieve different keyword sequences.
[0059] The DNA crawler architecture proposed in this invention exhibits good parallelism in complex information processing. The specific experimental results are as follows: like Figure 5 As shown, red and blue represent two different keywords. By embedding one or more keywords into the 118 bp DNA information chain and adding multiple corresponding sgRNAs to the reaction system simultaneously, the dCas9 complex follows a search mechanism combining one-dimensional sliding and three-dimensional collision in the liquid phase to achieve parallel retrieval of massive DNA libraries.
[0060] In step S1, when constructing the DNA crawler unit, the total reaction volume is 25 μL, specifically including: adding 0.625 μL of two types of sgRNA (total sgRNA concentration of 3 μM), 1 μL of dCas9 (625 nM), 2.5 μL of 10× 3.1 buffer, and 18.75 μL of ultrapure water to a 50 μL EP tube, and premixing thoroughly; then, incubating at 28°C on the top of the BioRAD PCR instrument and 25°C for 30 min to allow dCas9 and sgRNA to fully bind and form a binary complex; In step S2, 1.5 μL of dsDNA information strand with a concentration of 250 nM is added to the reaction system containing DNA crawler units.
[0061] In step S3, continue to maintain at 25 °C for 30 minutes to achieve binding incubation, complete PAM navigator recognition, target keyword matching, and ternary complex (dCas9-sgRNA-DNA) assembly.
[0062] In step S4, the DNA information strand carrying the target keyword is physically separated from the background pool in the form of a dCas9-sgRNA-DNA ternary complex by electrophoretic mobility shift analysis (EMSA).
[0063] The sample processing method for electrophoretic analysis is as follows: Take 25 μL of the reacted DNA sample, add 2.5 μL of 10× DNA Loading buffer, mix well by pipetting and centrifugation. For sample loading, add 20 μL of the sample to be tested and 2 μL of marker to each well. Perform electrophoresis on ice to increase resolution. Electrophoresis parameters: set the voltage to 60 V and run the gel at constant voltage for 120 minutes. After electrophoresis, place the gel under a UV gel imaging system and use the imaging system to take gel images to read the dsDNA information strand climbing status.
[0064] like Figure 6 As shown in the electrophoresis results, red and blue "1" represent DNA information strands carrying only one keyword's corresponding nucleotide sequence, thus binding to only one DNA crawler unit. Orange "2" represent DNA information strands carrying two keywords' corresponding nucleotide sequences, thus binding to two different DNA crawler units. Each DNA crawler unit has an independent sgRNA sequence. EMSA (electrophoretic mobility shift analysis) clearly demonstrates that this multi-target identification process can achieve specific binding to target files in the data pool. For a single target file, its migration position is positively correlated with the keyword matching strength, exhibiting a clear hierarchical distribution, indicating that target DNA carrying keywords can be separated from the system by electrophoresis.
[0065] Building upon this quantification capability, to further enhance the system's encoding capacity and feature representation ability, a combinatorial number system is introduced to achieve structured encoding of keywords. By randomly selecting k=2 keywords from a codebook of n=213 highly orthogonal keywords, theoretically 22577 unique combinations can be generated.
[0066] Compared with existing technologies, the DNA storage data inverted index retrieval method of the present invention has the following characteristics: 1) Rapid location in large-scale complex search spaces. This invention uses a DNA crawler architecture implemented in the CRISPR / dCas9 system to directly crawl keyword information, thereby introducing an inverted index mechanism and establishing a direct mapping between "keywords" and "documents," achieving an improvement from traditional addressing to content addressing. Based on the "DNA crawler" architecture, the retrieval complexity is optimized to O(log n).
[0067] 2) Supports fine-grained multi-mode retrieval. This invention supports parallel detection of extremely short keywords (< 20 bp) at the single-chain level, taking into account both the screening constraints of PAM gating and the high specificity of precise matching, giving the system an excellent response range.
[0068] 3) Supports parallel operation of reservoirs. By transforming complex computational search logic into programmable biochemical behavior, this invention enables the direct execution of parallel logic in a single reaction at the molecular level. This significantly reduces the overhead of intermediate steps in heterogeneous data processing, increases the query density of large-scale data, and expands the query scope.
[0069] 4) Lossless information reading and multi-level information feature preservation. This invention demonstrates its ability to preserve the statistical features of the original sample during the retrieval process, maintaining the inter-chain statistical distribution characteristics of the original sample before and after retrieval without introducing new base errors, thus ensuring the reusability of the data.
[0070] While specific embodiments of the present invention have been described above, they are not intended to limit the scope of the invention. They are merely illustrative examples. Other changes can be made to these embodiments without departing from the essence of the invention. Therefore, the scope of protection of the present invention is defined by the appended claims.
Claims
1. A method for retrieving DNA storage data using an inverted index, characterized in that, include: Step S0: Construct at least one DNA information strand, including a nucleotide sequence obtained by converting at least one keyword and a PAM motif at the end used to distinguish whether the DNA information strand is compliant; Step S1: Construct a DNA crawler unit based on the target keywords. The DNA crawler unit includes a dCas9 protein without endonuclease activity and a guide nucleic acid chain. The PI domain of the dCas9 protein serves as a PAM navigator, and the guide nucleic acid chain serves as the query statement for the keywords. Step S2: During the index search, the DNA information strand is mixed with the DNA crawler unit; Step S3: Use the PAM navigator to determine whether the end of the DNA information chain has a compliant PAM motif. If so, use a query statement to match the nucleotide sequence corresponding to the target keyword to achieve targeted search for the target keyword. Step S4: Using the physical or chemical characteristics of changes in DNA crawler units after binding to nucleotide sequences, the DNA information chain carrying the target keyword is separated.
2. The DNA storage data inverted index retrieval method according to claim 1, characterized in that, The guide nucleic acid strand is a programmable sgRNA, crRNA, or gRNA.
3. The DNA storage data inverted index retrieval method according to claim 1, characterized in that, The dCas9 protein has a REC domain, which is used to clamp a successfully matched nucleotide sequence corresponding to the target keyword and a guide nucleic acid strand.
4. The DNA storage data inverted index retrieval method according to claim 1, characterized in that, In step S0, multiple DNA information strands carrying different keywords coexist in the same oligonucleotide pool; In step S1, the number of DNA crawler units is one; or, the number of DNA crawler units is multiple, each DNA crawler unit having the same type of dCas9 protein and guide nucleic acid chain, while different DNA crawler units have different guide nucleic acid chains.
5. The DNA storage data inverted index retrieval method according to claim 4, characterized in that, In step S4, when the number of DNA crawler units is multiple, the separated DNA information chain with target keywords is a DNA information chain with multiple target keywords.
6. The DNA storage data inverted index retrieval method according to claim 1, characterized in that, The dCas9 protein is one of SpCas9 protein, VQR-Cas9 protein, VRER-Cas9 protein, and SpRY-Cas9 protein.
7. The DNA storage data inverted index retrieval method according to claim 1, characterized in that, The physical characteristics include at least one of molecular weight, spatial conformation, and charge, and the chemical characteristics include specific affinity.
8. The DNA storage data inverted index retrieval method according to claim 1, characterized in that, The DNA information strands carrying the target keywords are separated, specifically by physically separating the DNA information strands carrying the target keywords from the background pool under isothermal conditions through electrophoretic mobility variation analysis or magnetic bead sorting.
9. The DNA storage data inverted index retrieval method according to claim 2, characterized in that, The guide nucleic acid chain uses programmable sgRNA, and the molar ratio of dCas9 protein to sgRNA is 1:1 to 1:10.