Fig. 3
From: In silico prediction of lncRNA function using tissue specific and evolutionary conserved expression
Schematic workflow of our annotation algorithm. The correlation rank (CR) among tissue- or species-specific expression profiles is used to generate complete weighted single-tissue or single-species gene networks (STN or SSN). Previously known functional annotations linked to human genes are then used along with our gene networks to compute a functional prediction score (FPS) between each gene and every annotation term; in the case of SSN it is necessary to consider homology relations between human genes and those of the considered species. The information is then combined using logistic models. The models gives a list of predictions as output; each one consists of a score associated to a gene id and an ontology term