De novo assembled contigs

Trinity predicted denovo transcripts. The FASTQ sequence reads were assembled using Trinity which is specifically designed for <i>de novo </i>assembly of transcriptomes. Trinity was run on the paired end sequences with the fixed default k-mer size of 25 and minimum contig length of 200. Assembly algorithm runs on three steps, Inchworm, Chrysalis and Butterfly. Inchworm first assembles overlapping sequences using a greedy extension and reports the unique portions of alternatively spliced transcripts (contigs). Chrysalis then clusters related contigs into components (i.e., compXXXX_c0) and models complexity using de Bruijn graphs for each of the clusters. Finally, Butterfly processes the graphs and reports subcomponents (i.e., compXXXX_c0_seq1, compXXXX_c0_seq2, etc.), roughly corresponding to genes that are made up of individual transcripts.

本数据集为基于Trinity软件预测的从头（de novo）组装转录本。原始FASTQ测序读段通过Trinity软件完成组装，该工具专为转录组的从头组装（de novo assembly）研发。本次分析采用双端测序序列运行Trinity，设置固定的默认k聚体（k-mer）长度为25，且最小重叠群（contig）长度为200。该组装算法包含三个核心步骤：Inchworm、Chrysalis与Butterfly。首先，Inchworm通过贪心延伸策略组装重叠序列，并输出可变剪接转录本的独特片段，即重叠群。随后，Chrysalis将相关的重叠群聚类为组分（格式为compXXXX_c0），并为每个聚类使用德布鲁因图（de Bruijn graph）构建复杂度模型。最终，Butterfly对德布鲁因图进行处理，并输出子组分（格式为compXXXX_c0_seq1、compXXXX_c0_seq2等），这些子组分大致对应由单个转录本构成的基因。