MicroRNA Memory II: A Novel Scoring Integration Model for Prediction of Human Disease by MicroRNA/MicroRNA Quantum Multi-Interaction

Aims: To further investigate quantum characters of the microRNA (miRNA) gene as the disease memory device, we calculated neo-score tool for miRNA/miRNA multi-interaction. Since the potential miRNA/target interaction is not one-to-one correspondence; therefore, the network of miRNA/mRNA is too busy to achieve the goal of prognosis and diagnose human disease. Methodology: Neo-score tool based on quantum and electrodynamics for miRNA/miRNA multiinteraction, which are Dynamic Nexus Score (DNS) and Electric Field Tangent score (EFTS) is compare with Context+ Score Change (CSC) on PolymiRTS database, an integrated platform of the functional impact of genetic polymorphisms in miRNA seed regions and miRNA target sites within 149 human disease. Original Research Article Osone et al.; JAMPS, 5(3): 1-18, 2016; Article no.JAMPS.22095 2 Results: The DNS was correlated with CSC. Since the EFTS was mathematically functioned into the DNS, the function of the DNS to the EFTS was integrated together disease prediction on CSC. Further, a possibility was suggested in this context that miRNA/miRNA multi-interaction on the algorithmic function could be applied for specific discernment of disease without miRNA/target interaction complex. Conclusion: MiRNA/miRNA multi-interaction may have an important role for prognostic and diagnostic technologies for human health. This is the first report demonstrating that miRNA memory would manage the etiologies of human diseases.


INTRODUCTION
Recently, the function of non-coding RNAs (ncRNAs) has been investigated intensely, because ncRNAs have a key role in almost all cellular processes [1][2][3]. The ncRNAs are classified according to their length and function. The operation of long ncRNAs, whose length is more than 200 nucleotides, remains unclear. However, one type of small ncRNAs, the microRNAs (miRNAs), have an important direct or indirect role in the post-transcriptional and occasionally transcriptional regulation of gene expression and are highly conserved in various species. The miRNAs are approximately 22 nucleotides long, and their function of downregulating messenger RNAs (mRNAs) expression is predicted on the basis of their seed region, which consists of nucleotides 2-8 at the 5' end [4][5][6]. Competitive binding of the seed, target and competing endogenous RNAs regulates the protein expression level [7][8][9].
Various databases and miRNA target prediction tools are currently available. For instance, Target Scan identifies candidate miRNAs based on the strength of interaction between the seed region and target site of miRNAs. TargetScan introduced the Context+ Score value, which enables the selection of the most favorable target sites for miRNAs. Context+ Score value evaluates miRNA binding in the context of the entire 3'untranslated region (3'UTR) of the mRNA by summing contributions made by individual sites that are perfectly complementary to the miRNA seed [10]. Context+ Score value consists of the following six contributions: Site-type, 3' pairing, local AU nucleotides, position, target site abundance (TA) and seed-pairing stability (SPS) [10]. This value is calculated for one miRNA and one target region. Another database, miRBase (http://mirbase.org/), can be used to obtain mature miRNA sequences [11], and dbSNP collects data about human single nucleotide polymorphisms (SNPs) [12]. Because Target Scan, miRBase and dbSNP are only available independently to investigate miRNA/mRNA interactions and SNPs, a combined prediction tool was lacking for more complex prediction tasks. In 2014, Bhattacharya et al. integrated these three databases creating the PolymiRTS combined prediction tool [13][14][15]. PolymiRTS database 3.0 (PolymiRTS: http://compbio.uthsc.edu/miRSNP/) collects SNPs located in miRNAs and miRNA target sites that are verified experimentally. In addition, PolymiRTS contains numerous miRNA targets, which are implicated in human disease, and data about the interaction between miRNA seed regions and 3'UTRs of mRNAs. Context+ Score change (CSC) value is the Context+ Score value corrected with the effect of polymorphisms in the miRNA target sites and miRNA seed regions. Compared with references, a more negative CSC value indicates an increase in the binding strength of miRNA and mRNA. A specific mRNA is often targeted by the same miRNA at multiple sites. CSC represents the change in binding strength between multiple miRNAs and SNPs within the mRNAs.
RNA Wave 2000 Model is the model interpreted RNA information genes (Rigs) [16]. The RNA Wave 2000 Model has four pillars; 1) the Rigs as a mobile genetic element induce transcriptional and post-transcriptional silencing via networkingarchitecture; 2) the Rigs expand into the environmental cycle of life; 3) the Rigs can selfproliferate; 4) the Rigs have two types of information as resident and genomic ones. The RNA Wave 2000 Model describes a novel mechanism of miRNA regulation, which can be affected by environmental factors [17]. The coherent on the quantum direction in the high dimensions has been shown in the matrix with the miRNA qubit code; however, the coherent was unavailable to apply for challenging against miRNA/miRNA multi-interaction.
Previously, we developed from single miRNA/miRNA interaction to two new scores of represented multiple miRNA/miRNA interactions based on the RNA Wave 2000 Model. One is based on quantum [16] and the other is based on electrodynamics [18]. Quantum based score, Dynamic Nexus Score (DNS), and the electrodynamics based score, Electric Field Tangent score (EFTS), were compared with CSC because miRNA/miRNA interactions were not calculated by CSC. CSC was linearly correlated with DNS. DNS and EFTS were non-linearly associated with human disease independent of CSC. After gaining an insight into miRNA/miRNA multiple interactions using network prediction tools, we might use this information for tailored medical care.

MiRNA and mRNA Databases
All CSC data placed in Disease and Traits were prepared from PolymiRTS. In PolymiRTS, the cut off value of <50 of miRNA in one target mRNA was used. The mature.fa file in miRBase was downloaded in Release 21: June 2014, miRNA counts: 28,645 entries from a downloading page. All 2,588 of human miRNAs were extracted. CSCs related with 2,297 mRNA of 149 disease are used in this study (Table 1).

The New Score of DNS and DNSC
The miR-ket code and single miRNA/miRNA superposition in matrix were described previously [17]. The DNS is equal to the sum of each mathematical clause of the direct product when two miRNA sequences were diagnosed as a vector. Therefore, the miRNA group was defined as A, number of miRNA in A as N, each element of A as ‫ܣ‬ {i = 0,1,2, … , N − 1}, ‫ܣ‬ 's length as ‫ܯ‬ and sequence of each element as ‫ܣ‬ {j = 0,1,2, … ,23}.
The DNS (1,0) is shown as synergy between among miRNAs in our previous report. Next, δ DNS were calculated as a change of the DNS. When miRNA group element shown as ‫ܣ‬ ௫ and sequence of ‫ܣ‬ ௫ as ‫ܣ‬ ௫ {j = 0,1,2, … ,23} , δ DNS was expressed as follows, This equation was rearranged as follows (1,1).

The New Score of EFTS
Although the DNSC is the score that quantized the quantum energy in each base of miRNA, the new score based on electrodynamics was further developed. The electric field vector as E1 in any coordinate (x, y) to receive from electric charge Q in coordinate (X, Y) was expressed as follows.
When the number of electric charge increases, the electric field vector is demanded in the sum of the vector like the following expression as E2. The total number as N, each electric charge as Q ୧ , each coordinate as (X ୧ , Y ୧ ){i = 0,1,2, … , N − 1} was used as follows.

Fig. 1. EFTS related atoms in the nucleotide
In this formula, miRNA sequence was circularized in this model and (x, y) was any optional point.
The X ingredient (E x ) and the Y (E y ) ingredient of the electric field vector were separated as follows.

EFTS related atoms in the nucleotide
sequence was circularized in this model and (x, y) was any y ) ingredient of the electric field vector were separated as follows.
, each element of ‫ܧ‬ ௫ ሬሬሬሬԦ and ,2, … , N − 1} , respectively. Total electric field vector E4 was (2,5) The radius of miRNA was defined as 50 units and one unit was corresponding to a lattice. The electric field was expressed at lattice points. The value when the electric field vector in each coordinates was defined as Electric Field Tangent score (EFTS). The formula is as follows.
Further, EFTS 1 (2, 6) in a miRNA were changed to EFTS 2 in multiple miRNAs and the number of miRNA was shown as N.

Mathematical Functioning EFTSC into DNSC
From (1,0), the scalar product is corresponding to number of G, therefore, DNSC can be replaced for number of G. EFTSC was calculated by Fourier transformation as analogy of quantum computation. Mean of EFTSC, y and number of G (DNSC), x were functioned as follows,

Molecular Modeling of miRNA/miRNA Multi-Interaction
DNS and CSC analysis were performed to evaluate miRNA/miRNA multi-interaction in Agerelated disease, bone and muscle disease, cardiovascular disease, endocrine disease, eye disease, genital disease, gastrointestinal disturbance, immune disease, infection, lifestylerelated disease, neurodegeneration, renal failure, respiratory disease, skin disease, tooth disease and tumors (Table S1). The number of significant correlation between DNS and CSC was shown in Table 1. The relations between DNS from quantum scoring of miRNA/miRNA interaction and CSC from the context of miRNA/the target seed of mRNA were investigated (Fig. 2). DNSCs (x-axis) and CSCs (y-axis) were plotted in the case of a RFX7 target mRNA for chronic lymphocytic leukemia ( Fig. 2A). The means of CSCs and the values of DNSC were represented in the same panel of A and the correlation efficiency was calculated (Fig, 2B). R 2 was contained as 0.8488, therefore the average of CSCs was highly correlated to DNSCs. For all target mRNAs in all disease of this study, all correlation efficiencies were calculated between DNSCs and CSCs, and then P-values in Pearson correlation efficiency were represented (Fig. 2C). Gaussian distribution was observed. Under less than 1% of p-value were 391 (17%) of target, under less than 5% of p-value were 885 (39%) of target and under less than 10% of p-value were 1190 (52%) of target (Fig. 2D). Top 5 of linear relation (p<0.01) between DNSC and CSC were shown in Table 2. In age-related macular degeneration, SNP allele alteration of 3'UTR target of GLI3, MBL2 and ARMS2 were strongly correlated with DNSC. Significant correlations against truncating variants in SNPs of TXNDC16 transcript variant notv) 2, GRM7 tv1 and GREM1 tv1 were also observed in orofacial cleft. About coronary heart disease, three SNPs in HLA-DPA1 tv3, SLC22A3 and LAMC2 had the high relationship to DNSC. Crohn's disease was clearly implicated with SNPs of EHB1 tv3 and DNSC or PDLIM4 tv2 and DNSC. In the case of type 2 diabetes mellitus, PAX4, PLS1 tv2, MAP3K1 and CDC123 showed profound nexuses of DNSC. Further, DNS was allelespecifically correlated with SNPs of ST8SIA2, CREB5 tv3 and SRP72 in airflow obstruction, and CHRNA4 and IREB2 in chronic obstructive pulmonary disease SNPs of PCLO tv1, ESRRG tv4, STMN2 tv1 and GRM7 tv1 were significantly related with DNSC in neurodegeneration of the major category of human disease. These results indicate that miRNA/miRNA interaction computed by DNSs provides different functions for browsing and searching data about SNPs and human disease and would be important candidates for disease phenotype prediction and complex trait studies.

Relations among EFTSs, DNSCs and CSCs
When the electric field vector in EFTS was applied for nucleotides of miRNA, E3 vectors of each nucleotide as torus showed specific characters among miRNAs as example, miR-10a-5p, miR-27a-5p, miR-191-5p, miR-200a-5p, miR-573 and miR-592 (Fig. S2A). Further, under mathematically simulated condition, coherence state of two miRNAs, such as miR-27a-5p (stayed) and miR-573 (moving) was made as two scalars and E4 was presented as a torus in Hilbert space (Fig. S2B). Superposing state of two miRNA/miRNA interaction was identical among every prohibiting set of miRNA/miRNA interaction, which also be corresponding to identical points of miRNAs between the CSC in the x-axis (red), DNSC in the y-axis (green) and the EFTSC in the z-axis (blue) on Fig. 3A. Therefore, these scalar values were transformed for multi miRNA/miRNA coherence as binary annotation in consideration of Watson-Crick and wobble base pairings, and then EFTSC score was computed. To further improve the accuracy of miRNA/miRNA multi-interaction EFTSC was integrated into the relation between CSC and DNSC in chronic lymphocytic leukemia, RFX7 target ( Fig. 3A and Fig. S3A). These data showed that CSCs have linear relation with DNSCs (See Pearson correlation coefficient in Table S1) but all CSCs were not related with EFTSCs. Increasing of target further computation was performed. A figure of three-dimensional plot of mRNA related to chronic lymphocytic leukemia, Next, in bladder cancer, breast cancer, chronic lymphocytic leukemia and colorectal cancer, the DNSC plotting was performed and DNSCs in each disease were shown in Fig. dimensional plot of mRNA related with Fig  (Fig. S3C), CSCs, DNSCs and EFTSCs were visualized. In the case of another disease, atopic dermatitis, bladder cancer, biliary obstruction and chronic pulmonary disease, the DNSCs were processed at the same way ( Fig. 3C and Fig.  S3D). All CSCs, all DNSCs and all EFTSCs did not show significant relation among them under summing up of targets or disease. Since we did not obtain the possible formula to make impact of miRNA/miRNA multi-interaction, DNSC was compared to EFTSC in chronic lymp leukemia, RFX7 target (Fig. 3D). The number of G-base was plotted to EFTSC, non regression analysis was performed (blue line) and non-liner relation was observed. By definition of DNSC, DNSC is uniquely determined number of G in miRNA. Therefore, we confirmed relation between EFTSC and DNS (by number of G in miRNA).

Function of EFTSCs by DNSCs
As by non-linear regression analysis with Fourier transformation, number of G (x) and EFTSC (y) 7 11 targets were visualized in web (Fig. S3B). Next, in bladder cancer, breast cancer, chronic lymphocytic leukemia and colorectal cancer, the DNSC plotting was performed and DNSCs in 3B. In threeelated with Fig. 3B  (Fig. S3C), CSCs, DNSCs and EFTSCs were visualized. In the case of another disease, atopic dermatitis, bladder cancer, biliary obstruction and chronic pulmonary disease, the DNSCs were processed at the same way ( Fig. 3C and Fig. All CSCs, all DNSCs and all EFTSCs did not show significant relation among them under summing up of targets or disease. Since we did not obtain the possible formula to make impact of interaction, DNSC was compared to EFTSC in chronic lymphocytic leukemia, RFX7 target (Fig. 3D). The number of base was plotted to EFTSC, non-linear regression analysis was performed (blue line) liner relation was observed. By definition of DNSC, DNSC is uniquely determined number fore, we confirmed relation between EFTSC and DNS (by number of G in

Function of EFTSCs by DNSCs
linear regression analysis with Fourier transformation, number of G (x) and EFTSC (y) had some formula, a formula was prepared to be functioned by DNSCs. About top 5 of linear relation between DNSC and CSC (Table 2), 4, 5 and 6 of G (x) were substitute for the formula (Table 3). In each number of G in a phenotype, the maximum (red) and the minimum (blue) values of average of EFTSCs were Although Pearson correlation efficient did not identify the value of average of EFTSCs, non linear curve was computed on the targets, APOBEC3A tv1, TMEM129 tv2, PSCA tv1, CCNE1, FGFR3 tv3, SLC14A1 tv1, TP63 tv5 and TP63 tv3 in the bladder cancer ( Fig  3, the maximum and the minimum values tremendously different among x=4, 5 and 6. Therefore, each value was adjusted from 100 to make it easy to compare. Next, the average of each non-linear data was calculated on the same target as described above in the bladder cancer (Fig. 4B). Similar processes were performed in bladder cancer, breast cancer, chronic lymphocytic leukemia and cancer (Fig. 4C). Further, in atopic dermatitis, biliary obstruction, bladder cancer and chronic obstructive pulmonary disease, non analysis was done (Fig. 4D). These results indicate that function of EFTSC by DNSC specifically showed disease phenotype modes.
Correlation between quantum score and the target seed 18, 2016;Article no.JAMPS.22095 had some formula, a formula was prepared to be functioned by DNSCs. About top 5 of linear relation between DNSC and CSC (Table 2), 4, 5 and 6 of G (x) were substitute for the formula (Table 3). In each number of G in a phenotype, the maximum (red) and the minimum (blue) values of average of EFTSCs were represented. Although Pearson correlation efficient did not identify the value of average of EFTSCs, nonlinear curve was computed on the targets, APOBEC3A tv1, TMEM129 tv2, PSCA tv1, CCNE1, FGFR3 tv3, SLC14A1 tv1, TP63 tv5 and ncer (Fig. 4A). In Table  3, the maximum and the minimum values were tremendously different among x=4, 5 and 6. Therefore, each value was adjusted from -100 to 100 to make it easy to compare. Next, the linear data was calculated same target as described above in the Similar processes were performed in bladder cancer, breast cancer, chronic lymphocytic leukemia and colorectal cancer (Fig. 4C). Further, in atopic dermatitis, cer and chronic obstructive pulmonary disease, non-linear analysis was done (Fig. 4D). These results indicate that function of EFTSC by DNSC specifically showed disease phenotype modes.

Fig. 5. Scheme of relation between CSC, DNSC and EFTSC
Genome-wide association studies (GWAS) have identified miRNA and SNPs of target implicated in complex human traits and disease, therefore the same non-linear regression analysis as above was performed in cardiovascular disease Top 5 target SNPs, HLA-DPA1 tv3, SLC22A3, LAMC2 tv1, ABO (Venous thromboembolism; yellow) and ABO (Coronary heart disease; blue) ( Fig. 4E). Other 11 phenotypes, Age disease (A), bone and muscle (B), gastrointestinal disturbance (C), immune disease (D), infection (E), lifestyle-related disease (F), neurodegeneration (G), renal failure (H), respiratory disease (I), skin disease (J) and tumors (K), were also shown in Fig. S4(A These data show that EFTSCs was related with DNSCs in non-linear regression and miRNA/miRNA multi-interaction on this algorithmic function could specifically discern among each disease phenotype from one to another without CSCs.

DISCUSSION
Recently, miRNA regulatory networks and miRNA-related SNPs have been studied intensely to understand their role in human disease [19,20]. A GWAS was conducted to reveal associations between miRNAs and disease [21]. In this study, we used PolymiRTS, which contains information about miRNAs and polymorphisms in their targets. PolymiRTS connects numerous disease with S miRNAs, pri-miRNAs, miRNA promoters and miRNA targets [13]. SNPs in the miRNA genes 13

Fig. 5. Scheme of relation between CSC, DNSC and EFTSC
wide association studies (GWAS) have identified miRNA and SNPs of target implicated in complex human traits and disease, therefore linear regression analysis as above was performed in cardiovascular disease v3, SLC22A3, LAMC2 tv1, ABO (Venous thromboembolism; yellow) and ABO (Coronary heart disease; blue) ( Fig. 4E). Other 11 phenotypes, Age-related disease (A), bone and muscle (B), (C), immune disease related disease (F), neurodegeneration (G), renal failure (H), respiratory disease (I), skin disease (J) and tumors (K), were also shown in Fig. S4(A-K). These data show that EFTSCs was related with linear regression and teraction on this algorithmic function could specifically discern among each disease phenotype from one to Recently, miRNA regulatory networks and related SNPs have been studied intensely to understand their role in human ]. A GWAS was conducted to reveal associations between miRNAs and ]. In this study, we used PolymiRTS, which contains information about miRNAs and polymorphisms in their targets. PolymiRTS connects numerous disease with SNPs in pre-miRNAs, miRNA promoters and ]. SNPs in the miRNA genes and in their 3'UTR mRNA targets are largely diverse among individuals, and cause various phenotypes and disease. In our model experiments, we used the PolymiRTS database, because according to GWAS, SNPs in target sites are associated with disease. Role of these genetic traits has been confirmed using luciferase-based gain-and loss assays in vitro [19,22], by comparative analysis of the human genome, and a network experiment has recently revealed the target miRNAs (miR-943 and miR-571) for the therapy of hepatocellular carcinoma [23]. Furthermore, based on a functional assay using a miRNA lentivector library, the tamoxifen responsive miRNA has been identified as target in breast cancer using the genome-integrated pre forced expression assay [24]. This suggests that the role of SNPs in the 3'UTR are identified with gain-and loss-of-function models in humans, and multiple parameters of miRNAs and their targets are essential in the diagnosis, treatment and prognosis of disease. In addition, SNPs in miRNA targets might predict miRNA function; therefore, pharmacogenomics and molecular epidemiology can utilize these results. GWAS revealed disease-related genes, which are related or not related to various biological pathways and mechanisms. We hypothesized that the latter case could be explained with miRNA/miRNA interactions, because almost of all biological events are controlled by miRNAs. For instance, there is a strong association between DNS and CSC values of GRM7 of Parkinson's disease. GRM7 can modulate 18, 2016;Article no.JAMPS.22095 and in their 3'UTR mRNA targets are largely diverse among individuals, and cause various In our model experiments, we used the PolymiRTS database, because according to GWAS, SNPs in target sites are associated with disease. Role of these genetic traits has been confirmed using and loss-of-function ], by comparative analysis of the human genome, and a network-based experiment has recently revealed the target 571) for the therapy ]. Furthermore, assay using a miRNA lentivector library, the tamoxifen responsive miRNA has been identified as target in breast integrated pre-miRNA ]. This suggests that the role of SNPs in the 3'UTR are identified with function models in humans, and and their targets are essential in the diagnosis, treatment and prognosis of disease. In addition, SNPs in miRNA targets might predict miRNA function; therefore, pharmacogenomics and molecular epidemiology can utilize these results. GWAS related genes, which are related or not related to various biological pathways and mechanisms. We hypothesized that the latter case could be explained with miRNA/miRNA interactions, because almost of all biological events are controlled by miRNAs.
tance, there is a strong association between DNS and CSC values of GRM7 of Parkinson's disease. GRM7 can modulate neurotransmitter release and neuronal excitability. Furthermore, GRM7 controls neural differentiation via the phosphorylation of CREB and regulation of YAP expression [25 according to GWAS, GRM7 is linked to bipolar [26], depression [27], hyperactive disorder schizophrenia [29] and age-related hearing loss [30]. On the contrary, ESRRG is related to neurodegeneration; however, according to the literature, ESRRG is mainly implicated in the energy-balance of the whole body function of ESRRG in neurotransmission remains unknown. In the case of this GWAS, our data shows that ESRRG might function via miRNA/miRNA multiple interactions. Thus, our findings suggest that miRNA/miRNA multiple interactions are crucial in human disease and sufficient for the prediction of diagnosis and prognosis (see Fig. 6).
Torus RNA could be advantageous for biological prevalence because it can serve as template for rolling circle replication and be stable due to be resistant to exonucleases [32 experiments, it can be simulated that torus model of miRNA can transmit genetic information among miRNAs therefore, the circular type of RNA might have an important role for its function 14 neurotransmitter release and neuronal excitability. Furthermore, GRM7 controls neural differentiation via the phosphorylation of CREB [25]. Therefore, GWAS, GRM7 is linked to bipolar ], hyperactive disorder [28], related hearing loss On the contrary, ESRRG is related to neurodegeneration; however, according to the literature, ESRRG is mainly implicated in the balance of the whole body [31], and function of ESRRG in neurotransmission remains GWAS, our data shows that ESRRG might function via miRNA/miRNA multiple interactions. Thus, our findings suggest that miRNA/miRNA multiple interactions are crucial in human disease and sufficient for the prediction of diagnosis and Torus RNA could be advantageous for biological prevalence because it can serve as template for rolling circle replication and be stable due to be [32]. In our experiments, it can be simulated that torus model can transmit genetic information among miRNAs therefore, the circular type of RNA might have an important role for its function to architect the mechanisms of tuning of miRNA. Quite recently, it has been reported that natural circular RNA were identified in human cells and the circular RNA (1,400 nts) can bind to miR the sponge [33]. Further Y RNA is known as a family of non-coding and a natural circular RNA [34], and human Y3 and Y5 (25 nts each) were annotated as miR-1979 and miR miRBase, respectively. Artificially, the repetitive circular RNAs (approximate 14 nts) are catalyzed by polymerase from ribozymes and 34 nts RNA were made as circular RNA by using a template directed non-enzymatic ligation results indicated that natural mobile torus miRNA could be bio-generated and have a f transmit gene information. Further, circular RNA has been related with cancer [36 above miR-1979 and miR-1975 are circulating miRNA, and as biomarkers, miR associated with multiple sclerosis chronic congestive heart failure [38 miR-1975 is implicated in liver cancer suggesting that the circulating circular genes would have some function to induce human disease, therefore, DNSC and EFTS scorings under tours condition are suitable for human disease prediction to be relevant to GWAS. Fig. 6. The future prospecting model 18, 2016;Article no.JAMPS.22095 to architect the mechanisms of tuning of miRNA. Quite recently, it has been reported that natural in human cells and the circular RNA (1,400 nts) can bind to miR-7 as ]. Further Y RNA is known as a natural circular RNA ], and human Y3 and Y5 (25 nts each) were 1979 and miR-1975 in ctively. Artificially, the repetitive circular RNAs (approximate 14 nts) are catalyzed by polymerase from ribozymes and 34 nts RNA were made as circular RNA by using a templateenzymatic ligation [35]. These results indicated that natural mobile torus miRNA generated and have a function to transmit gene information. Further, circular RNA [36]. Intriguingly, 1975 are circulating miRNA, and as biomarkers, miR-1979 is associated with multiple sclerosis [37] and [38]. In addition, 1975 is implicated in liver cancer [39], suggesting that the circulating circular miRNA genes would have some function to induce human disease, therefore, DNSC and EFTS scorings under tours condition are suitable for human disease prediction to be relevant to We classified human disease with algorithms based on electrodynamics and quantum energy of miRNA nucleotides. EFTS based on electrodynamics corresponds to nucleotide changes in miRNAs, such as keto-enol tautomerism [40], changes induced by environmental stress, including ultraviolet rays [41] and A/I RNA editing [42]. On the contrary, DNS corresponds to the presence of G bases in miRNAs but not tautomerism. Guanine generate G-quartets [43] or scale of the fish caused by changes in the magnetic field [44]. In addition, miRNA SNPs changed to guanine are more effective [45]. Accordingly, G-dependent miRNA/miRNA interactions, based on the miRket code, are important factors in the association of miRNAs and human disease [46]. While CSC and EFTS were not related, CSC correlated with DNS. Therefore, guanine bases in miRNAs are often associated with disease-related polymorphisms. Furthermore, in the targets, approximately 40% of DNS values have a linear relationship with CSC. Therefore, DNS might be a useful tool to remove false positive data from a large miRNA dataset in targets.
As a miRNA contain approximately 22 nucleotides, the DNS value can only have 276 varieties. However, corresponding to Watson-Crick base pair and wobble base pair, the EFTS value can have 2 131 +2 65 varieties. We found that EFTS convergently has a non-linear relationship with DNS. The non-linear regression analysis with a Fourier transformation showed that each disease can be represented with a specific nonlinear curve. Because in classical mechanics all physical events can be described with specific wave properties using quantum mechanics, our results would be rational to physical criteria. Thus, EFTS (related to classical mechanics) and DNS (related to quantum mechanics) had a nonlinear relationship, suggesting that each disease is visualized using specific wave properties. In addition, using non-linear curves without CSC, phenotype of human disease can be identified. Therefore, human disease may be predicted by the miRNA/miRNA interaction as miRNA memory only.
As the calculation of EFTS and DNS values is still problematic, further research is necessary. Firstly, when we set EFTS, atoms that form a Watson-Crick base pair [47] or wobble base pair [48] were selected as active ones. However, nucleic acids can also form a Hoogsteen base pair [49]. Hoogsteen base pairs are related to the G-quartet structure and Fragile X mental retardation protein (FMRP) nucleic acid chaperone has facilitated miRNA assembly inviting recognition of G-quartet structures on target mRNAs [50]; therefore, EFTS calculations have to be set up for Hoogsteen base pairs. Secondly, although we presented the relationship between miRNA/miRNA interactions and disease based on GWAS, the expression levels of miRNAs were not considered, because miRNA-mRNA binding sites were identified with direct mapping experiments such as cross linking, ligation and sequencing of hybrids (CLASH) [51]. Therefore, integration of comparative miRNA qualitative expression levels might improve the performance of the model. Further progression was necessary for precision scoring.

CONCLUSION
Neo-scores based on the quantum and electrodynamic data of the miRNA/miRNA interaction, DNS and EFTS, respectively, is shown in Fig. 5. DNS was correlated with CSC using PolymiRTS data to evaluate the relationship between miRNAs and protein coding genes invade in 149 human disease. Because EFTS was mathematically functioned into DNS, function of DNS to EFTS was integrated for disease prediction on CSC. Furthermore, in this context, miRNA/miRNA interactions in the algorithmic function might be applied for specific discernment of disease with the miRNA/target interaction complex. Thus, miRNA/miRNA interactions as miRNA disease memory may have an important role in the a priori disease control of human health. Utility of neo scores is shown in Fig. 6. Artificial Intelligence (AI) combined with the EFTS/DNSbased prediction method might help in the diagnosis of disease. From a technical point of view, our method might improve the prediction of miRNA/miRNA interactions with AI by using more accurate information.

CONSENT
It is not applicable.

ETHICAL APPROVAL
It is not applicable.

Supplemental Information includes Extended Experimental
Procedures, seven figures, and three tables and can be found