Results Seed developmental stage specific library construction, sequencing and sequence analysis In higher plants, most miRNAs regulate their targets via cleavage, which normally occurs Ruxolitinib between the tenth and eleventh nucleotides of the complementary region be tween the miRNA and the mRNA target. The 3 cleavage fragments contain both a free 5 monopho sphate and a 3 polyA tail. So, these cleavage products can be successfully ligated with RNA ligase, whereas full length cDNAs with a 5 cap structure or other RNAs lacking the 5 monophosphate group are not compatible for ligation and thus will be unavailable for subse quent amplification and sequencing reactions. Five different degradome libraries, which capture the cleaved mRNAs, were constructed from cotyledons and seed coats from different seed developmental stages.

These represented early maturation seed and mid maturation, the stages when the biosynthetic capacity of the seed is maximal and proteins and oils are accumulated at a high rate. In addition, we constructed a library from the yellow cotyledons that are undergoing dehydration and desiccation. SBS sequencing of these libraries produced raw reads from 10 million to 45 million. After removal of low quality sequences and adapter removal, 95% of these reads lengths were 20 or 21 nt in length as expected from the cloning procedure. More than 97% of reads mapped to the soybean genome available at the Phyto zome data base. We also used the computationally predicted cDNA transcripts from the soybean genome sequence consisting of 78,773 high and low confidence gene models for mapping degradome reads and found that more than 95% of reads matched to the Glyma models.

These data indicate our degra dome libraries to be of high quality and efficiency in recovering degraded mRNA targets that should contain the sequence profile resulting from miRNA directed cleavage. Systematic identification of miRNA targets in soybean Systematic identification of miRNA targets was accom plished using previously described methods by analyzing the 20 and 21 nt reads with the CleaveLand pipeline for miRNA target identification using all Glycine max miRNAs from miRBase. The identified targets were grouped into five categories by the program based on the relative abundance of the number of reads map ping to the predicted miRNA target site relative to other sites in the gene model.

Those in category 0 clearly have the majority of tags located at the miRNA guided cleavage site. The identified Dacomitinib miRNA targets using degradome se quencing are presented in the form of target plots that plot the abundance of the signatures rela tive to their position in the transcript. Representa tive t plots are shown, one from each of four different degradome libraries such as cotyledon, seed coat, cotyledon and seed coat.