PhyloFold: Precise and Swift Prediction of RNA Secondary Structures to Incorporate Phylogeny among Homologs

Masaki Tagashira

A large quantity of functional ncRNAs have been discovered through sequencing technology, including high throughput sequencing (Maxam and Gilbert, 1977; Bentley et al., 2008) . The RNAs are related to various biological processes such as epigenetic silencing (Pasmant et al., 2011) , splicing regulation (Ji et al., 2003) , translational control (Long and Caceres, 2009 ), apoptosis regulation, and cell cycle control (Kino et al., 2010) . However, the functions of a large number of RNAs have not yet been uncovered. Secondary structures of the RNAs are an important key to uncovering the functions of the RNAs because the structures play critical roles in biological processes, (Wu et al., 1991) and secondary structures are often conserved among homologs even in the case where the sequences of the homologs are nonconserved. (Klein and Eddy, 2003) To more correctly predict secondary structures, simultaneous consideration of sequence alignment and secondary structure, or structural alignment (Sankoff, 1985; Hofacker et al., 2004; Havgaard et al., 2007) , has been taken into account. (Hamada et al., 2009c (Hamada et al., , 2011 Tan et al., 2017) However, the consideration has been proven intractable (Sankoff, 1985) and thus has been conducted roughly to decompose structural alignments. (Hamada et al., 2009c,a; Sato et al., 2012; Tan et al., 2017) To more accurately accomplish this goal, the PhyloFold method ( Figure 1 ) was developed in this study. The method predicts secondary structures of homologs considering likely pairwise structual alignments between the homologs by a sparsification technique. (1) The method computes pairwise structural alignment probabilities for each pair of input homologs.