Prediction Of Noncoding RNA Secondary Structures And Detection Of Structured RNAs

Higher organisms up to more than half of the DNA is transcribed into RNA,the vast majority of which is non-coding RNA,These RNAs are not translated into proteins and are capable of exercising their biological functions at the RNA level?Although non-coding RNA is now more and more important,most of the known non-coding RNAs sequences are short and structurally unstable,it is difficult to detect their secondary structures,while the secondary structures determine the function of non-coding RNA in cells,so a special algorithm is needed to predict its secondary structure in order to better study the relationship between structure and function.In all probes that detect RNA,the single-stranded regions is often selected as the target sequence,however,some functional RNAs are highly structurally.Therefore,there is a need for a method capable of detecting structured RNA in order to real-time quantitative and detection of its function.In this paper,we study the prediction of non-coding RNA secondary structure and the detection of structured RNA,and study the following aspects:Prediction of Small RNA Secondary Structure Based on Reverse Complementary Folding Method.Small non-coding RNA is now a more mature study,especially microRNA.Although the sequence of small RNA is short and the structure is unstable,the structure of small RNA has some conservatism.And the existing algorithms of RNA prediction basically did not consider the structural characteristics of small RNA.Aiming at these problems,we improved the method of predicting the small RNA secondary structure,and proposed a reverse complementary folding method to predict the secondary structure of small RNA.First,according to the small RNA(especially pre-microRNA)secondary structure without multi-branch loop,the elimination of multi-branch loop calculation;then,the binary path matrix is established at the same time when the matching matrix is established.Backtracking in the path matrix can avoid the formation of multi-branch loop and self-folding.Finally,the free energy can be added to all the predicted secondary structures,and finding the structure with the minimum free energy is the predicted structure by the method in this paper.The sensitivity,specificity and MCC values were used as the evaluation criteria,and the data in the pre-microRNA data and the PDB database were used as the test data.Compared with the RNAfold and RNAstructure methods,the method of this paper is more accurate.Prediction of non-coding RNA secondary structure based on replacement of stems.The secondary structure of RNA can be regarded as the combination of stem and loop,and the appropriate and compatible combinations of stem and stem can be obtained from the stem pool.The connection between the stem and the stem is the loop region in the secondary structure.Based on the above theory,this paper proposes a method for prediction of non-coding RNA secondary structure based on replacement of stems--RSRNA methods.The algorithm to establish the stem pool,so that the stem area to a certain degree of sorting,making the calculation process,only the need for local search can find the best stem;Then the new method is used to calculate the compatibility of stem and stem,and the compatible stems are combined into a secondary structure and the free energy is calculated;after comparing the free energy of these structures,select the stem;as the stem as a fixed structure is not changing,the rest of the stems all re-selection;repeat all of the above steps to predict the second stem and all the stems until the predicted structure is incompatible with all stems and ultimately predict the RNA secondary structure.This method can be well compatible with the nested structure in the RNA secondary structure,and the randomness of the algorithm becomes smaller,and the each performance of algorithm is much higher than that of the heuristic algorithm and the minimum free energy algorithm.The results show that the RSRNA algorithm proposed in this paper is superior to the commonly used minimum free energy algorithm,such as RNAfold,Srna,RNAstructure and Mfold method.Preparation of highly structured RNA molecules.tRNALys3 has a typical clover-like secondary structure and inverted L-shaped tertiary structure,which can be predicted by the RSRNA method.Human tRNALys3 is closely related to human immunodeficiency virus type 1(HIV-1)and other lentiviruses.These viruses use tRNALys3 as a primer for reverse transcription.So we use human tRNALys3 molecule as an example to detect.Before experimentally detecting the molecule,it is necessary to use the experimental method to produce a highly structured tRNALys3 molecule.The DNA primers are designed,the template DNA is prepared using primers,and then the template DNA is used to transcribe the tRNALys3 molecule.In order to obtain a high purity RNA solution,not only the conventional solution purification method was used,but also the electrophoretic cleavage and purification method was used in the purification of the tRNA solution.According to the experimental phenomena and experimental results,RNA and impurities in the gel plate can be very clear separation,obtained a higher purity RNA solution.Absorbance detection and RNA validation results are satisfactory,fully meet the experimental expectations and requirements.Detection of highly structured small non-coding RNA.The use of molecular beacons is a versatile method to detect RNAs.Typically,a single-stranded region of RNA is selected as a target sequence for molecular beacons.Therefore,the detection of highly structured short RNAs,such as tRNAs,seems to be difficult.In this study,as an example of highly structured target RNA,we used human tRNALys3.No long single-stranded region more than 8 nt is present in this tRNA,which is much shorter than the standard target length of molecular beacons(20 nt).Therefore,it is necessary to use computational methods to predict the secondary structure of MB and the change of structure and energy before and after MB binding to target RNA.In addition,we designed two types of MBs:(i)MBs in which the anti-target sequence started from the 50 stem and continued to the 30 end of the loop of the MB,and(ii)MBs in which the anti-target sequence was completely included in the loop.By using these MBs,we investigated which design was optimal for detecting highly structured RNA and attempted to identify the best target region.For comparison with the highly structured tRNA,short target RNAs that were unstructured or less-structured were also used in the RNA detection experiments.This study showed that sensitive detection of highly structured RNAs using molecular beacons was much more difficult than that of unstructured or less structured RNAs.However,efficient detection of the tRNALys3 was achieved by selecting the best target region,i.e.,the region around the D arm,probably due to the ease of unfolding of this arm.Accordingly,our findings suggested that molecular beacons may have applications in the detection of highly structured RNAs related to various biological functions and diseases.