Annotation and classification of LTR retrotransposon elements

We performed de novo searches for LTR retrotransposons against the eight rice genome sequences using LTR_STRUC (McCarthy and McDonald 2003). False positives caused by long tandem repeats were manually removed by BLAST searches. All intact LTR retrotransposons were classified into Ty1-copia, Ty3-gypsy, and unclassified groups according to the order of ORFs using PFAM (Finn et al. 2008). The RT sequences were retrieved from each retrotransposon element and further checked by homology searches against the published RT genes available from GyDB (http://gydb.org/) (Llorens et al. 2011). They were aligned using ClustalW (Larkin et al. 2007) and manually curated (Table 1). Previous LTR retrotransposon family nomenclature (see Figure 1 and Figure 2) was determined using BLAST searches with LTR retrotransposons downloaded from TIGR (Ouyang and Buell 2004) and Repbase (Jurka 1998, 2000; Jurka et al. 2005). A homology search of the genome sequence was performed using RepeatMasker (Smit et al. 1996–2010). All intact LTR retrotransposon sequences generated by LTR_STRUC (McCarthy and McDonald 2003) were classified into families using BLASTClust (http://www.ncbi.nlm.nih.gov/Web/Newsltr/Spring04/blastlab.html) and all-to-all BLAST of 5′-LTR sequences, followed by manual inspection (Llorens et al. 2011). The family classification standard was considered acceptable if >50% of the 5′-LTRs and sequence identity was >80%. Detailed information of LTR retrotransposon families is provided in Supplemental Material, Table S1 in File S1.

SAT, O. sativa ssp. japonica. cv. Nipponbare (Release 7); RUF, O. rufipogon; NIV, O. nivara; GLA, O. glaberrima; BAR, O. barthii; GLU, O. glumaepatula; LON, O. longistaminata; MER, O. meridionalis; LTR, long terminal repeat.

Insertion times and read depth analysis of the top 23 LTR retrotransposon families. (A) Insertion times of the exemplar LTR retrotransposons from SAT. Black circles indicate mean values, black bars signify 25–75% of values, dark gray lines represent 5–95% of values, and green circles denote extreme values. The light gray horizontal lines show divergence times between SAT and the other seven species. Those inserted earlier than 5 MY are set at ∼5 MY. Ty3-gypsy (red dots) Ty1-copia (blue dots) elements are distinguished. (B) Heatmap of the proportions of LTR read depth compared within each family across the eight rice genomes. BAR, O. barthii; GLA, O. glaberrima; GLU, O. glumaepatula; LON, O. longistaminata; LTR, long terminal repeat; MER, O. meridionalis; MY, million years; Mya, million years ago; NIV, O. nivara; OAL, Oryza AA genome of LTR retrotransposons; RUF, O. rufipogon; SAT, O. sativa ssp. japonica. cv. Nipponbare (Release 7).

Phylogenetic trees of representative LTR retrotransposon lineages across the eight AA-genome Oryza species. Neighbor joining and unrooted trees were constructed based on sequences of RT genes for Ty1-copia (A) and Ty3-gypsy (B). The backbone of the RT trees and retroelements from SAT are shown in black, while colored branches indicate those from other seven species. LTR retrotransposons from NIV and RUF (the closest to SAT), GLA and BAR (similar divergence time to SAT), GLU, LON, and MER are colored in red, orange, green, light blue, and dark blue, respectively. BAR, O. barthii; GLA, O. glaberrima; GLU, O. glumaepatula; LON, O. longistaminata; LTR, long terminal repeat; MER, O. meridionalis; MY, million years; Mya, million years ago; NIV, O. nivara; OAL, Oryza AA genome of LTR retrotransposons; RT, reverse transcriptase; RUF, O. rufipogon; SAT, O. sativa ssp. japonica. cv. Nipponbare (Release 7).