Background: Metazoans transcribe many long non-coding RNAs (lncRNAs) that are poorly conserved and whose function remains unknown. This has raised the questions of what fraction of the predicted lncRNAs is actually functional, and whether selection can effectively constrain lncRNAs in species with small effective population sizes such as human populations. Results: Here we evaluate signatures of selection in human lncRNAs using inter-specific data and intra-specific comparisons from five major populations, as well as by assessing relationships between sequence variation and predictions of secondary structure. In all analyses we included a reference of functionally characterized lncRNAs. Altogether, our results show compelling evidence of recent purifying selection acting on both characterized and predicted lncRNAs. We found that RNA secondary structure constrains sequence variation in lncRNAs, so that polymorphisms are depleted in paired regions with low accessibility and tend to be neutral with respect to structural stability. Conclusions: Important implications of our results are that secondary structure plays a role in the functionality of lncRNAs, and that the set of predicted lncRNAs contains a large fraction of functional ones that may play key roles that remain to be discovered.

Toni Gabaldon's (TG’s) group acknowledges support from the Spanish Ministry of Economy and Competitiveness grants “Centro de Excelencia Severo Ochoa 2013-2017” SEV-2012-0208 and BFU2015-67107, cofounded by the European Regional Development Fund (ERDF); from the European Union and ERC Seventh Framework Programme (FP7/2007-2013) under grant agreements FP7-PEOPLE-2013-ITN-606786 and ERC-2012-StG-310325; and a grant from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement number H2020-MSCA-ITN-2014-642095.