پديد آورندگان :
عباسي، وحيده دانشگاه فردوسي مشهد - دانشكده كشاورزي - گروه علوم دامي , نصيري، محمدرضا دانشگاه فردوسي مشهد - دانشكده كشاورزي - گروه علوم دامي , جوادمنش، علي دانشگاه فردوسي مشهد - دانشكده كشاورزي - گروه علوم دامي
كليدواژه :
بيوانفورماتيك , روش از آغاز , ريزRNAها , گوسفند
چكيده فارسي :
در سالهاي اخير با روشن شدن اهميت ريزRNAها در فرآيندهاي حياتي بدن موجودات زنده، شناسايي ژنهاي ريزRNA اهميت زيادي پيدا كردهاست از سويي ديگر روشهاي بيوانفورماتيكي نيز گسترش يافته و سبب شده تا روند شناسايي ريزRNAها با سرعت بيشتر و هزينه ي كمتري پيش برود. تحقيقات در سالهاي اخير بر روي گونه هاي زيادي از جمله انسان و موش و همچنين حيوانات اهلي از قبيل گاو و بز و مرغ باعث شده تا در اين گونه ها ريزRNAهاي زيادي گزارش شود با اين حال در گونه ي گوسفند هنوز اطلاعات نسبتا كمي وجود دارد. در اين مطالعه پس از پيش بيني ريزRNAهاي كروموزوم 20 گوسفند، جهت تاييد با اورتولوگها در ساير گونه ها مقايسه و در نهايت براي بررسي بيان از داده هاي ترنسكريپتوم گوسفند استفاده شد همچنين در اين تحقيق براي ارزيابي روش مورد استفاده از داده هاي شاهد مثبت استفاده شد. با استفاده از اين روش در كل 400 ريز RNA و از اين تعداد 81 ريزRNA جديد در كروموزوم 20 گوسفند پيش بيني شدند. بررسي ترنسكريپتوم نشان داد كه از اين تعداد 33 ريزRNA فقط در بافت عضله، 10 ريزRNA فقط در بافت كبد، 35 ريزRNA در هر دو بافت كبد و ماهيچه و 3 ريزRNA نيز در مخلوط بافتهاي قلب، كليه، مغز، تخمدان، پوست، چربي سفيد و ريه بيان شدند. مقايسه با پايگاه داده miRbase مشخص كرد كه تعداد 64 ريزRNA از ريزRNAهاي پيدا شده براي اولين بار گزارش شدند. حساسيت و دقت انتخاب اين روش برابر 67 و 95 درصد ارزيابي شد و از اين رو ميتوان گفت اين روش ميتواند به عنوان مكمل روشهاي پيش بيني ريزRNA اطلاعات ارزشمندي را فراهم كند. از اين روش مي توان براي پيش بيني ريزRNAها در ساير كروموزوم هاي گوسفند و يا در ساير پستانداران استفاده نمود.
چكيده لاتين :
Introduction MicroRNAs (miRNAs) are small noncoding RNA molecules that are found in plants, animals and some
viruses and play important role in regulation of transcription (3). Recently, importance of miRNA roles in biology of
living organisms has been discovered, thus miRNAs identification became more important (1). Experimental detection
of miRNAs can be obtained using different miRNA profiling methods, such as quantitative real-time PCR, microarray,
and high-throughput RNA sequencing technologies. Since the identification and verification of miRNA by laboratory
methods is time-consuming and costly (4), for improving miRNA identification and lowering costs, it is more
reasonable that miRNA loci predicted by reliable bioinformatic approaches then experimental methods used for
confirmation. This may decrease false positive results. Recently, several hundreds of miRNAs reported in variety of
species including human and mouse, as well as domestic animals such as cattle, pig, chicken and goat, however there
are relatively less information about sheep miRNAs. In this study, we developed a method for prediction and in silico
validation of miRNAs located on ovine chromosome twenty.
Materials and Methods In this study, an ab initio approach was used. The sequence of ovine chromosome 20 was used
as input for EMBOSS and Sequence-Structure Motif Base: Pre-miRNA Prediction Webservers applications, then all
predicted stem and loop structures were entered into mfold software. Pre-miRNA features for them were calculated and
sequences that had these features were selected. Since the probability of miRNA presence in the coding region is very
low, miRNAs that were predicted in the coding regions were removed. To confirm the prediction of miRNAs, selected
sequences were homology searched within all registered miRNAs in miRBase. In order to evaluate the in silico
expression of miRNAs, predicted miRNAs were BLASTed against expression data from Sequence Read Archive (SRA)
of ovine muscle tissue (Accession: PRJNA223213), liver tissue (Accession: GSM1366318) and mixture of tissues
including heart, kidney, brain, liver, ovary, lung, skin, and adipose (Accession: GSE56643). In order to evaluate the
accuracy of this method, a positive control region including a cluster of validated miRNAs from ovine chromosome 18
were analyzed by the same method and sensitivity and selectivity of this method were calculated based on this region
from chromosome 18.
Results and Discussion After prediction by softwares and investigation of pre-miRNAs features by mfold, 400 stem
and loop structures that had pre-miRNA features were chosen. Fifty miRNAs from those miRNAs contained conserved
mature miRNAs sequence and 350 of them were recognized as novel miRNAs based on registered miRNAs in the
mirBase. None of the novel miRNAs were located in the coding regions. In silico validation of these novel miRNAs in
SRA data was indicated that 81 miRNAs are expressed in different ovine tissues including 33 in muscle and 10 in liver.
Results on the positive control data showed that 40 miRNAs were predicted which majority of them (36 miRNAs) were
already validated by experiments. This indicates a high reliability for this method. With putting in sensitivity and
selectivity formulas, both of two factors were calculated and it was observed that the sensitivity and selectivity values
for our method were 67% and 90%, respectively. Fewer studies accomplished in detection of ovine miRNAs in compare
to other farm animals. In previous studies to identify miRNAs in ovine species, mostly laboratory-based or comparative
methods were used. This was the first study that used SRA database to check miRNA expression in RNAseq data in
order to decrease the discovery of false positive results. Comparing this method with others including CID-miRNA
(19), miRPara (20), VMir (6) and miRNAFold (18) methods, we may conclude that this method’s sensitivity is less than
CID-miRNA, miRPara, miRNAFold and srnaloop. Although selectivity for this method is higher than all above methods because the false positive in this method is less than other method. This method showed high selectivity and
low FP that due to improved prediction method for identify miRNAs.
Conclusion In the current study, predicted ovine miRNAs were validated by an in silico method using SRA database
that resulted in a higher reliability than other ab initio approaches. Although this method is not very fast and fully
automated. With running this method on ovine chromosome 20, 81 novel miRNA were predicted which were expressed
in different tissues of sheep. This method could be applied on other ovine chromosomes as well as other mammalian
species, although future validation by experimental approaches is required.
