Mito-nuclear phylogenetic discordance in Bivalvia is well known. In particular, the monophyly of Amarsipobranchia (Heterodonta + Pteriomorphia), retrieved from mitochondrial markers, contrasts with the monophyly of Heteroconchia (Heterodonta + Palaeoheterodonta), retrieved from nuclear markers. However, since oxidative phosphorylation nuclear markers support the Amarsipobranchia hypothesis instead of the Heteroconchia one, interacting subunits of the mitochondrial complexes ought to share the same phylogenetic signal notwithstanding the genomic source, which is different from the signal obtained from other nuclear markers. This may be a clue of coevolution between nuclear and mitochondrial genes. In this work we inferred the phylogenetic signal from mitochondrial and nuclear oxidative phosphorylation markers exploiting different phylogenetic approaches and added two more datasets for comparison: genes of the glycolytic pathway and genes related to the biogenesis of regulative small noncoding RNAs. All trees inferred from mitochondrial and nuclear subunits of the mitochondrial complexes support the monophyly of Amarsipobranchia, regardless of the phylogenetic pipeline. However, not every single marker agrees with this topology: this is clearly visible in nuclear subunits that do not directly interact with the mitochondrial counterparts. Overall, our data support the hypothesis of a coevolution between nuclear and mitochondrial genes for the oxidative phosphorylation. Moreover, we suggest a relationship between mitochondrial topology and different nucleotide composition between clades, which could be associated to the highly variable gene arrangement in Bivalvia.