**BinPacker**

**Trans-AbySS**

**Oases**

usage among all assemblers.

**IDBA-Tran**

and number of recovered full-length transcripts [31].

Trans-abyss is a method and pipeline for the collection and analysis of short transcriptomic data. Abyss assembly process consists of single-ended and double-ended stages. The singleended stage is also based on the *de Bruijn* graph structure; when parameter k is given, it is transformed into tiled k-mer represented as read nodes and (k-1) bases are superimposed as directed edges. Allelic differences, minor changes in the sequence and repetitive random base invocation errors lead to 'bubbles' throughout the graph. Once these errors have been removed in the k-mer space, the single-ended contigs defined by the 'walk' clear across the graph. In the matched tier phase, the pairs aligned in the single-ended contigs define the empirical distribution of the distances of the pairs. Single-ended readings of different contigs to the co-aligned pairs and empirical distribution then intercontig distance and combined to form contigs are paired end contigs that can be combined [23]. Trans-AbySS reaches the end by creating direct sequenced readings with Bruijn graphics, removing possible errors from the middle and solving each connected Bruijn graph for each connected component. Compared to other assembler programs the lowest percentage of chimera is seen in Trans-AbySS [30]. Comparative studies showed that with Trinity, Trans-ABySS performed best in gene coverage

60 Applications of RNA-Seq and Omics Strategies - From Microorganisms to Human Health

Oases is a RNA transcriptome assembler that contains many developmental constructs. Combines multiple k-mers and topological analysis methods. In addition, it uses the dynamic error correction feature developed for RNA-Seq data. Assembly process of Oases takes place by creating independent assemblies, which vary according to the length of the k-mers, and then assembling them all together in one assembly. In each assembly, readings are used to generate *de Bruijn*, and then faults are simplified, organized into a scaffold, divided into loci and eventually analysed. Then dynamic correction is performed and Oases creates contigs sets of clusters called loci. Since it is more likely to be unique, long contigs treated first when the scaffold is constructed and faults that may arise from alternative splices are eliminated. Oases provide a robust pipeline from RNA-Seq readings to generate full-length assemblies of transcripts. Especially designed for dealing with RNA-Seq condition, unequal coverage and alternative spliced situations [24]. Oases-Velvet produced the highest number of chimeric transcripts at different k-mer sizes and it has the highest RAM (i.e. random access memory)

IDBA-Tran uses a different approach. Firstly, it produces small *de Bruijn* graphs and enlarges the graph with larger k values. Subsequently, transcripts are found on a large Bruijn graph, BinPacker reshapes the problems and generates full-length transcripts by following the aggregated graph line generated by various techniques used in Bridger. Some advantages of BinPacker: (i) BinPacker allows the use of user-defined k-mer values for best performance and (ii) BinPacker uses a strict mathematical model. This allows the BinPacker to achieve a lower false positive rate at the same sensitivity level. (iii) BinPacker makes full use of the step depth applied to graphics, so that the assembly results are more accurate. BinPacker combines transcripts on every merging graph it creates [26]. BinPacker is more unsuccessful than other programs on chimeric data [31].
