The majority of the human genome does not code for proteins. However, today it is accepted that this noncoding genome regulates important aspects in the cell. Several families of non-coding RNAs have been described, but the most prevalent one is the so-called long non-coding RNAs (lncRNAs). LncRNAs are a type of non-coding transcripts with essential regulatory functions in a large variety of cellular processes, many of which are mis-regulated in human diseases including cancer and neurological disorders. Still, lncRNAs’ function and their role in disease is largely enigmatic. Single-molecule techniques such as cryo-EM, atomic force microscopy and optical/magnetic tweezers are revolutionizing our understanding of how molecules and cells function, but the lncRNA field has not yet benefited from this revolution due to the complexity of lncRNA biology. Thousands of long transcripts comprise a few structured regions with functional meaning, embedded within large and flexible regions of a RNA molecule. Furthermore, the same RNA domain can bind several proteins with completely different functional consequences, whilst several proteins can bind the same site of a lncRNA. Thus, we hypothesize that the structure but not the sequence is the key in understanding their roles. However, any structural and mechanistic study depends on predicting the RNA domains that are functionally relevant. This project aims to decrypt lncRNA function in humans on a transcriptome-wide scale, by bringing together computational modelling, lncRNA cell biology, cryo-EM and biophysical methods. Computational modelling will predict and classify the complex lncRNA interactome. Single-molecule methods and experiments in cells will relate major structural groups with mechanisms of action. We expect that the global integration of experimental structural and functional information, computational predictions, and interactions of human lncRNAs will provide a paradigm shift in the lncRNA field.