四款miRNA鉴定工具在非模式植物中应用的系统比较  

A Systematic Comparison of Four miRNA Identification Tools for Application in Non-model Plants

作　　者：朱楚萌 朱安丹 Zhu Chumeng;Zhu Andan(The Germplasm Bank of Wild Species,Kunming Institute of Botany,Chinese Academy of Sciences,Kunming,650201;College of Life Sciences,University of Chinese Academy of Sciences,Beijing,100049)

机构地区：[1]中国科学院昆明植物研究所,中国西南野生生物种质资源库,昆明650201 [2]中国科学院大学生命科学学院,北京100049

出　　处：《分子植物育种》2023年第10期3297-3307,共11页Molecular Plant Breeding

基　　金：国家自然科学基金项目(No.31860534)资助。

摘　　要：为评估四款主流的miRNA鉴定工具(miRDeep-P2,Mirnovo,ShortStack,UEA s RNA workbench)在非模式植物中应用的有效性和普适性,本研究选取模式植物拟南芥(Arabidopsis thaliana)以及非模式植物玉米(Zea mays)和智利草莓(Fragaria chiloensis)3个物种的小RNA高通量数据作为测试数据集,从用户体验、运行性能、结果评估、参数设置等方面对这四款工具进行了系统比较。结果表明,就用户体验而言,miRDeep-P2和ShortStack在总体易用性上优于Mirnovo和UEA s RNA workbench;就运行性能而言,miRDeep-P2和UEA s RNA workbench运行较高效;经过从头鉴定和同源性注释之后的结果比较发现,miRDeep-P2和ShortStack在不同物种间的miRNA鉴定结果整体好于另外两款工具,表明这两款工具可能也适用于其他非模式植物;另外,从评估指标来看,ShortStack鉴定的未注释的miRNAs数量最低,表明其可能具有较低假阳性,而miRDeep-P2具有最高的敏感性和准确性。最后选择不同的比对参数(错配数为0,1,2)对miRDeep-P2和ShortStack的鉴定结果进行比较,结果表明即使在不同的错配数设置下,ShortStack的鉴定结果在种内较为一致,而miRDeep-P2鉴定的miRNAs总数变化范围较大。总体而言,miRDeep-P2在综合表现上优于其他三款工具,但推荐在鉴定非模式植物的miRNAs时最好结合其他证据过滤假阳性,以及必要时对鉴定的miRNAs结构进行手动检查。本研究可为从事相关领域的研究者在工具选择和参数设置上提供一定参考。To evaluate the efficiency and universality of four popular miRNA identification tools(miRDeep-P2,Mirnovo,ShortStack,and UEA sRNA workbench)for application in non-model plants,three small RNA(sRNA)sequencing datasets from the model plant Arabidopsis thaliana,non-model plants Zea mays and Fragaria chiloensis were selected as the test datasets,and the four tools were systematically compared in terms of user experience,operational performance,result evaluation,and parameter settings.The results showed that,for the user experience,miRDeep-P2 and ShortStack generally performed better in usability than Mirnovo and UEA sRNA workbench;while miDeep-P2 and UEA sRNA workbench could run effectively in computational time;after de novo identification and homologous annotation,it was found that miRDeep-P2 and ShortStack performed overall better across different species than the other two tools,suggesting that the two tools may also be applicable to other non-model plants;moreover,ShortStack identified the minimum unannotated miRNAs among these tools,indicating the lowest false positives,and miRDeep-P2 had the highest sensitivity and accuracy.Finally,the identification results of miRDeep-P2 and ShortStack were compared by choosing different mapping parameters(mismatch number with 0,1,2),which showed that the identification results of ShortStack were more consistent within species even under different mismatch number settings,while the total number of miRNAs identified by miRDeep-P2 had a wider range of variation.In a word,miRDeep-P2 outperformed the other three tools in terms of overall performance,but filtering false positives with other evidence to identify miRNAs from non-model plants is highly recommended,as well as manual inspection of the structure of identified miRNAs if necessary.This study can provide some reference for researchers working in related fields in terms of tool selection and parameter settings.

关 键 词：microRNA鉴定 高通量小RNA测序 软件评估 非模式植物 

分 类 号：Q943.2[生物学—植物学]