Evidence of higher order patterns in information transmission between nucleotide sequences and folded molecular shapes of RNA

Abstract

This contribution is a short version of a full paper submitted to Bioinformatics. DNA data transcribe into single stranded RNA, which folds into specific configurations. On the level of contact structures these are described by RNA secondary structures. Here we stipulate that RNA structures provide semantics for sequential DNA data. Accordingly we study the correlation between RNA sequences and RNA structures. We compute the partition function of sequences with respect to a fixed structure. We present a Boltzmann sampler and obtain the a priori probability of specific sequence patterns in such samples. We present a detailed analysis for the two PDB-structures, 2JXV (hairpin) and 2N3R (3-branch multi-loop). We localize where specific sequence patterns occur, contrast the energy spectrum of Boltzmann sampled sequences versus those sequences that refold into the same structure and derive a criterion to identify native structures.