A General Computational Framework for Sequence-Level Thermodynamic Analysis of Modeled RNA–DNA and DNA–DNA Duplex States: A Proof-of-Concept Study in HIV-1

Source abstract

Abstract We present a sequence-level computational framework designed to be applicable across viral genomes and demonstrate it in a proof-of-concept analysis of HIV-1. Sixteen predefined regions of the HXB2 reference genome (K03455.1) were evaluated using two RNAstructure 6.6 procedures: OligoWalk for an RNA-target/DNA-oligo duplex calculation and DuplexFold for a DNA-DNA bimolecular minimum-free-energy reference. We define a model-specific contrast, ΔΔG_model = ΔG_RD − ΔG_DD, rather than a universal same-algorithm ΔΔG. Fifteen of 16 regions had negative contrasts; the panel mean was −8.00 kcal/mol (SD 5.45). A two-sided one-sample test against zero gave t₁₅ = −5.87, p = 1.54×10⁻⁵. The directional result persisted in overlap-sensitivity analyses excluding TAR and/or R_5LTR. These results constitute a sequence-level, testable computational prediction under the specified modeling procedures. They do not establish cellular R-loop formation, genomic integration, or a causal biological mechanism. The study is limited to one reference genome and to 16 non-independent, partly overlapping regions.