High-throughput sequencing technology led significant advances in functional genomics, giving the opportunity to pay particular attention to the role of specific biological entities. Recently, researchers focused on long non-coding RNAs (lncRNAs), i.e. transcripts that are longer than 200 nucleotides which are not transcribed into proteins. The main motivation comes from their influence on the development of human diseases. However, known relationships between lncRNAs and diseases are still poor and their in-lab validation is still expensive. In this paper, we propose a computational approach, based on heterogeneous clustering, which is able to predict possibly unknown lncRNA-disease relationships by analyzing complex heterogeneous networks consisting of several interacting biological entities of different types. The proposed method exploits overlapping and hierarchically organized heterogeneous clusters, which are able to catch multiple roles of lncRNAs and diseases at different levels of granularity. Our experimental evaluation, performed on a heterogeneous network consisting of microRNAs, lncRNAs, diseases, genes and their known relationships, shows that the proposed method is able to obtain better results with respect to existing methods.
Identifying lncRNA-disease relationships via heterogeneous clustering
Barracchia, Emanuele Pio;Pio, Gianvito;Malerba, Donato;Ceci, Michelangelo
2018-01-01
Abstract
High-throughput sequencing technology led significant advances in functional genomics, giving the opportunity to pay particular attention to the role of specific biological entities. Recently, researchers focused on long non-coding RNAs (lncRNAs), i.e. transcripts that are longer than 200 nucleotides which are not transcribed into proteins. The main motivation comes from their influence on the development of human diseases. However, known relationships between lncRNAs and diseases are still poor and their in-lab validation is still expensive. In this paper, we propose a computational approach, based on heterogeneous clustering, which is able to predict possibly unknown lncRNA-disease relationships by analyzing complex heterogeneous networks consisting of several interacting biological entities of different types. The proposed method exploits overlapping and hierarchically organized heterogeneous clusters, which are able to catch multiple roles of lncRNAs and diseases at different levels of granularity. Our experimental evaluation, performed on a heterogeneous network consisting of microRNAs, lncRNAs, diseases, genes and their known relationships, shows that the proposed method is able to obtain better results with respect to existing methods.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.