TransPS: A Transcriptome Post Scaffolding Method for Assembling High Quality Contigs

Abstract

Motivation

As the development of the high throughput sequencing technologies, transcriptome can be sequenced with a low price and high efficiency. Sequence assembly approaches have been renewed to meet the new requirements from new sequencing technologies. Assembly strategies are important for biologists who need to assemble the transcriptome generated in their experiments. However, some modern de novo assembly strategies generate a large section of redundant contigs due to sequence variations, which greatly affect downstream analysis and experiments. This work proposed TransPS, a post transcriptome scaffolding method to generate high quality transcriptomes.

Results

TransPS shows promising results on the test transcriptome data sets where the redundancy is greatly reduced by at least 50%, while the coverage is improved considerately.

Availability

The web server and source code are available at https://bioinformatics.cs.vt.edu/zhanglab/transps/

1. Introduction

The rapid development of the next generation sequencing technologies is bringing new development of the genome assembly tools that are able to handle the features of the new technologies as well as the downstream applications based on the quality of the sequencing. The demand of low cost of sequencing has driven high-through and high efficiency modern sequencing technologies, such as 454, Illumina, etc. Despite of the advantages of the next generation sequencing technologies, the length of the sequence generated by these modern instruments is considerably short (~100bp), which brings challenge to sequence assembly algorithms. Similar to short read genome assembly, transcriptome assembly needs to connect short, low quality reads. However, transcriptome assembly is even more difficulty than genome assembly with regards to considering sequencing depth, strand specific or transcript variants (Martin et al., 2011).

There are three typical transcriptome assembly strategies, reference based strategy, de novo strategy or combined strategy that joins the two. Widely used transcriptome assembly tools include Cufflinks (Adam et al., 2011) for reference based assembly, and Trinity (Grabherr et al., 2011) and Oases (Marcel et al., 2012) for de novo assembly. De novo assembler generates large number of contigs due to repetitive elements. They are also very sensitive to sequencing errors, which result in big redundancy in the output contigs. Paired-read sequencing technology could help to reduce the amount of contigs as the known intermediate distance between read pairs can be used to place contigs in their likely order and orientation. Assembly tools like Trinity does not include a scaffolding step, while the majority provide a scaffolding option only as a built-in function which cannot be independently controlled or effectively reduce the amount of contigs.

Very few previous studies about independently scaffolding preassembled transcriptome. SSPACE (Boetzer et al., 2011) is a tool to scaffold pre-assembled genome contigs using paired end data. In this work, we developed TransPS (Transcriptome Post Scaffolding) programming to scaffold pre-assembled transcriptome from any desired assemblers by using a reference species. The feature of TransPS is easy to implement and efficiently remove redundancy in the original config set. TransPS reduced the size of original contigs by at least %50 for our test datasets, while the quality of the scaffolding contigs are greatly enhanced in terms of the coverages.

2. MATERIALS AND METHODS

2.1. Datasets Description

Six test datasets were used to validate the TransPS pipeline, four of which were taken from NCBI Transcriptome Shotgun Assembly (TSA) projects, including Dendroctonus frontalis (TSA record: GAFI01), Dendroctonus ponderosae (TSA record: GAFX01), Diaphorina citric (TSA record: GACJ01) and Ixodes ricinus (TSA record: GACI01). The other two datasets are Aedes albopictus and Aedes hensilli transcriptome, which are newly sequenced via 454 and Illumina RNA-seq by Virginia Bioinformatics Institute. Please see supplementary table S1 for the corresponding reference sequences and pre-assembled methods used for each organism and other details.

2.2 Post Scaffolding

The input of the TransPS are pre-assembled contigs and the search results from NCBI BLASTX programming (Altschul et al., 1997) by aligning the contigs to the reference amino acid sequences. The post scaffolding procedure follows an align-layout-consensus structure, consisting of three major stages. First, original contigs search best alignment from the reference amino acid sequences database using BLASTX programming. Second, contigs are placed in a right order and orientation in terms of the coordinates aligned to the reference. Third, non-redundant sequences matched to the same sequence of the reference are scaffolded into one contig. Supplementary Figure S1 shows an overview of the procedure.

In the alignment stage, BLASTX was used in order to use protein sequence to guide the alignment of DNA sequence by searching a protein database using a nucleotide query. The best target sequence (lowest evalue and highest bitscore) is selected as the best match with the query sequence. It is possible that multiple cintigs match with the same protein sequence due to incomplete assembly of sequences, which is referred as a one-to-many matching, otherwise a one-to-one matching. In the case of one-to-many matching, all the contigs matching with the same protein are connected and put into a set of sequences T_i, where i represents protein i with which all contigs in T_i match.

To produce a high quality set of contigs, it is important to remove the redundant sequences generated by the assemblers. The contigs in T_i are divided into three subsets, contigs that are accepted as a final contig (A_i) that needs no further processing, contigs that are redundant (R_i), and sequences that are used to scaffold into one contig (S_i). The one to-one matching contigs goes to set A. For oneto-many matching, sequences in T_i are assigned to R_i, A_i or S_i. If two contigs match with the same protein at the same region or they are overlapped with each other for a certain percentage (user defined argument), then one of the two contigs is a redundant and removed to R_i. After moving the redundant and accepted sequences, all the remaining sequences in T_i go to subset S_i for scaffolding. In the case of only one config left for scaffolding, it goes to A_i subset. This procedure is summarized in the supplementary Figure S1 (above the dashed line). In the following two stages, we focus on the contigs for scaffolding in S_i.

In the layout stage, the purpose is to order the contigs in S_i based on the reference coordinates provided by the alignment. Intuitively, to ensure the consistency with the reference genome, the contigs are sorted in terms of the matching start positions on the reference protein in a ascending order.

In the consensus stage, the purpose is to scaffold the original contigs in S_i from s₁ to s_n into one high quality contig based on the order obtained from the previous stage (s_i ⎿ S_i (1≤i≤n)). The most important part for scaffolding is to estimate the “scaffolding parts”, which are determined based on two cases, overlapped contigs or separated configs (see supplementary method for details).

The output of TransPS are contigs in three different groups, accepted contigs (A), scaffolded contigs (S) and unused contigs (R). The contigs in the redundant set (R) could be real redundancy due to genetic variants or different transcripts due to alternative splicing. A matching map between the original contigs and the corresponding reference protein sequences used in the scaffolding algorithm is also provided.

3. RESULTS

Table 3 shows the number of scaffolds from the original contigs that match with at least one reference sequence (No. matched original). The average number of configs per scaffold is 2 to 3. The contigs in the redundant set (R) account for most of the original contigs (% Shrinked). Please see supplementary Table S2 for detailed distribution of original contigs in three different groups.

The new scaffolded contigs are compared with the original transcriptome contigs by measuring the coverage ratio against the matched reference protein sequence. Coverage ratio was calculated as the matched percentage between the scaffolded (or original) contig and the matched reference protein sequence in a global sequence alignment. The global sequence alignment was implemented by NAP (Huang et al., 1996), a programming doing global sequence alignment between a amino acid sequence and a nucleotide sequence. The redundant and accepted sequences in the original transcriptome set were also removed for comparison. As shown in Figure 1, the scaffolded contigs have much higher coverage ratio compared with the original ones and the median of scaffolded group is significant higher than the original group (p < 2.2e–16 ) in all the tested datasets.

Fig. 1.

Coverage comparison between scaffolds and original contigs for different data sets

4. CONCLUSION

This paper proposed a referenced based post transcriptome scaffolding method that considerately reduced the size of the original de novo assembled contigs while the qualities are greatly improved with regarding to coverage ratio. As more and more genomes become available in the public databases, it is more likely than not that a species with a complete genome closely related to a genome we are interested in. In this case, a post assembling by taking the information in the existed genome provides advantages to better guide the assembling of the studied genome.

Table 1.

Number of scaffolds generated by TransPS from different datasets

	NO. matched original	NO. scaffolds	NO.contig /scaffold	%Shrinked
De. Frontal	15,095	496	2.17	70
De. Ponderosae	16,457	433	2.15	74
Di. citri	16,732	165	2.21	90
Ix. ricinus	7,787	71	2.04	57
Ae. albopictus	47,445	3,863	3.36	74
Ae. hensilli	62,088	2,651	2.73	80

NO.contig/scaffold: average number of original contigs per scaffold

% Shrinked: percentage of “redundant” contigs