Refining Mitochondrial Intron Classification With ERPIN: Identification Based on Conservation of Sequence Plus Secondary Structure Motifs

• Published in: Frontiers in microbiology Vol. 13; p. 866187
• Main Authors: Prince, Samuel, Munoz, Carl, Filion-Bienvenue, Fannie, Rioux, Pierre, Sarrasin, Matt, Lang, B Franz
• Format: Journal Article
• Language: English
• Published: Switzerland Frontiers Media S.A 18.03.2022
• Subjects: covariance models; ERPIN; group I; infernal; Microbiology; mitochondrial introns; RNA structure; infernal; mitochondrial introns; ERPIN; covariance models; group I; RNA structure
• ISSN: 1664-302X; 1664-302X
• DOI: 10.3389/fmicb.2022.866187

Mitochondrial genomes-in particular those of fungi-often encode genes with a large number of Group I and Group II introns that are conserved at both the sequence and the RNA structure level. They provide a rich resource for the investigation of intron and gene structure, self- and protein-guided splicing mechanisms, and intron evolution. Yet, the degree of sequence conservation of introns is limited, and the primary sequence differs considerably among the distinct intron sub-groups. It makes intron identification, classification, structural modeling, and the inference of gene models a most challenging and error-prone task-frequently passed on to an "expert" for manual intervention. To reduce the need for manual curation of intron structures and mitochondrial gene models, computational methods using ERPIN sequence profiles were initially developed in 2007. Here we present a refinement of search models and alignments using the now abundant publicly available fungal mtDNA sequences. In addition, we have tested in how far members of the originally proposed sub-groups are clearly distinguished and validated by our computational approach. We confirm clearly distinct mitochondrial Group I sub-groups IA1, IA3, IB3, IC1, IC2, and ID. Yet, IB1, IB2, and IB4 ERPIN models are overlapping substantially in predictions, and are therefore combined and reported as IB. We have further explored the conversion of our ERPIN profiles into covariance models (CM). Current limitations and prospects of the CM approach will be discussed.