基于Illumina RNA-Seq短序列的转录组从头组装软件比较与优化(英文)

Comparing De Novo Transcriptome Assemblers Using Illumina RNA-Seq Reads

利用公共数据库中果蝇F1代和栽培水稻基于高通量Illumina测序平台的RNA-Seq短序列数据,比较了8个(ABySS,Velvet,SOAPdenovo,Oases,Trinity,Multiple-k,T-IDBA and Trans-ABySS)转录组从头组装软件。结果显示,在基于单一k-mer和多重k-mer方法的两类软件中,Trinity和Trans-ABySS分别表现出最好的组装性能,而其它软件性能比较接近。我们还发现基于多重k-mer比单一k-mer可以组装获得更多的总碱基数目,但是即使利用最好的多重k-mer组装软件,所获得的数据质量也比研究人员所期望的要低。鉴于此,我们提出了"ETM"优化方法,将多重k-mer方法组合到Trinity中,使其在具有最好的组装性能的基础上兼具了多重k-mer的优势,测试结果显示了该方法具有一定的优越性。我们的研究结果为用户选择合适的软件提供了依据,对推动基于高通量Illumina测序的转录组研究具有重要意义。

In this study, we carried out a systematic comparison of the de novo transcriptome assembly performance of eight assemblers （ ABYSS, Velvet, SOAPdenovo, Oases, Trinity, Multiple-k, T-IDBA and Trans-ABySS）, processing Illumina RNA-Seq reads from F1 hybrids （ Drosophila MS） of Drosophila melanogaster and Drosophila sechellia and cultivated rice. Our study showed that Trinity and Trans-ABySS were the most effective for producing transcriptomes from our trial datasets using single k-mer and multiple k-mer methods, respectively, although the per- formance levels of the other tested assemblers were comparable. We found that using single k-mer assemblers generally produced fewer total numbers of bases than multiple k-met assemblers, although even the best assembler＇s results showed lower quality than some researchers may desire. Therefore, we developed and tested a novel de novo transeriptome assembly method, ETM, which employs a combination of multiple k-mer tools with Trinity assembler. The ETM method yielded superior results from our trial datasets. Our results will assist the growing number of transcriptome projects using Illumina RNA-Seq reads and provide guidelines for choosing appropriate software.