The Annotation And Comparison Of MITEs And LTR-RTs In 6 Citrus Species

Citrus, as the most important fruit crop in China, increasing the fruit yield and improving its fruit quality are the basic needs. A previous study reported that the insertion of a Copia-like LTR-RT upstream Ruby locus in sweet orange led to the formation of bloody orange, which showed potential application on quality improvement of TEs in Citrus. Previous TEs were annotated using different criteria and pipelines, which made comparison among different Citrus species impossible. We annotated the TEs in the 6 sequenced Citrus species, mainly using MITE-HUNTER, LTRFINDER and RepeatMasker with unified criteria, which revealed the sequence characteristics and evolution features of MITEs and LTR-RTs and provided some active TE families and whole genome annotation dataset for the following studies.We obtained 62,010 full-length MITEs from 110 distinguish families, and found that MITE-related sequences account for about 3% length of each genome. By analyzing the distribution of sequence divergence and phylogenetic trees of MITEs, we found that DTM63 may be an active family in some Citrus species, and this is the first discovered active Mutator-like MITE family in Citrus. By homology-search analysis, we inferred that DTM63 transpose via cross-mobilization mechanism. Meanwhile, we found that some ancient MITE families such as DTM53 had experienced transposition burst before the divergence of Citrus genus and Severinia genus. The insertion of MITEs led much polymorphic loci among the Citrus genomes, and the inter-genus ratio of P/A polymorphism is much bigger than intra-genus ratio. And 12.9% of the small RNAs which we got from a previous study derived from MITE in sweet orange, which suggested that some MITEs could participate in transcriptional regulation via MITE-derived small RNAs in Citrus.Besides, we got 5,983 Complete LTRs and 13,399 Solo LTRs, and found that LTR-related sequences account for 21.2%~33.2% length of each genome. By constructing the RT sequences phylogenetic trees of Complete LTRs, we found that there are members from the 12 lineages which had been present before the divergence of eudicot and monocot species in Citrus and different lineages showed different sequence characteristics and activation patterns. We found that the insertion time obey an exponential distribution, which indicated that there is a dynamic balance between the insertion and elimination of LTRs. By comparing the relative content of LTR-related sequences on each chromosome, we found that LTR-related sequences predominantly enrich in centromere-proximal regions. And we found there were some correlations between Solo LTR formation and some factors, such as LTR insertion time, LTR length and INTERNAL length, which indicated that those factors probably play roles on Solo LTR formation.