Uncovering full-length transcript isoforms of sugarcane cultivar Khon Kaen 3 using single-molecule long-read sequencing

Abstract

Background: Sugarcane is an important global food crop and energy resource. To facilitate the sugarcane improvement program, genome and gene information are important for studying traits at the molecular level. Most currently available transcriptome data for sugarcane were generated using second-generation sequencing platforms, which provide short reads. The de novo assembled transcripts from these data are limited in length, and hence may be incomplete and inaccurate, especially for long RNAs.

Methods: We generated a transcriptome dataset of leaf tissue from a commercial Thai sugarcane cultivar Khon Kaen 3 (KK3) using PacBio RS II single-molecule long-read sequencing by the Iso-Seq method. Short-read RNA-Seq data were generated from the same RNA sample using the Ion Proton platform for reducing base calling errors.

Results: A total of 119,339 error-corrected transcripts were generated with the N50 length of 3,611 bp, which is on average longer than any previously reported sugarcane transcriptome dataset. 110,253 sequences (92.4%) contain an open reading frame (ORF) of at least 300 bp long with ORF N50 of 1,416 bp. The mean lengths of 5' and 3' untranslated regions in 73,795 sequences with complete ORFs are 1,249 and 1,187 bp, respectively. 4,774 transcripts are putatively novel full-length transcripts which do not match with a previous Iso-Seq study of sugarcane. We annotated the functions of 68,962 putative full-length transcripts with at least 90% coverage when compared with homologous protein coding sequences in other plants.

Discussion: The new catalog of transcripts will be useful for genome annotation, identification of splicing variants, SNP identification, and other research pertaining to the sugarcane improvement program. The putatively novel transcripts suggest unique features of KK3, although more data from different tissues and stages of development are needed to establish a reference transcriptome of this cultivar.

Keywords: Full-length transcripts; Iso-Seq; KK3; Khon Kaen 3; PacBio sequencing; Single-molecule long-read sequencing; Sugarcane; Transcriptome.

Figure 1. Distribution of hit plant species from BLAST search of PacBio-isoforms.

Pie chart shows the fraction of hit plant species based on the best hit obtained from BLASTX search of PacBio transcripts against Phytozome plant proteins.

Figure 2. Comparison of BLAST hits from different sequence databases.

Venn diagram shows overlaps of BLAST analysis results among compared databases, namely sugarcane nucleotide, Phytozome plant protein, and NCBI nr protein databases.

Figure 3. Length distribution of predicted ORFs of PacBio-isoforms.

Frequency distribution graphs are displayed for (A) ORF length of complete ORFs (green) and partial ORFs (blue) and (B) UTR length calculated from the complete ORFs separated into 5′ UTR (green) and 3′ UTR (blue).

Figure 4. COG classification of sugarcane PacBio transcripts in comparison to sorghum transcripts.

The frequency distributions of transcripts assigned to each functional class of KOG database were displayed for sugarcane PacBio transcripts (black) and sorghum transcripts available from Phytozome database (blue).

Figure 5. Length comparison of PacBio transcripts and their matched sequences.

The graph shows the frequency distribution for the ratio of the length of PacBio transcript to its matched sequences from (A) sugarcane and (B) Phytozome plant transcript databases. The percentages of coverage on hit plant CDS sequence are shown in (C).

Figure 6. Length distribution of PacBio transcripts.

Distributions of sequence length are displayed for sugarcane PacBio transcripts generated in the present study (green) and in the Hoang et al. (2017) study (blue).

Figure 7. Alternative splicing patterns of PacBio transcripts.

SpliceGrapher diagrams illustrate the splicing patterns of transcripts compared among the sorghum transcript annotation, the matched PacBio transcripts from the Hoang et al. (2017) study, the matched PacBio transcripts generated in the present study, and the matched PacBio transcripts combined from both studies. (A) Peptidase S24/S26A/S26B/S26C family protein (Sobic.002G223200) and (B) sucrose-phosphatase (Sobic.004G151800). Each color represents type of splicing event according to the data label.