Annotating function to differentially expressed LincRNAs in myelodysplastic syndrome using a network-based method

Abstract

Motivation: Long non-coding RNAs (lncRNAs) have been implicated in the regulation of diverse biological functions. The number of newly identified lncRNAs has increased dramatically in recent years but their expression and function have not yet been described from most diseases. To elucidate lncRNA function in human disease, we have developed a novel network based method (NLCFA) integrating correlations between lncRNA, protein coding genes and noncoding miRNAs. We have also integrated target gene associations and protein-protein interactions and designed our model to provide information on the combined influence of mRNAs, lncRNAs and miRNAs on cellular signal transduction networks.

Results: We have generated lncRNA expression profiles from the CD34+ haematopoietic stem and progenitor cells (HSPCs) from patients with Myelodysplastic syndromes (MDS) and healthy donors. We report, for the first time, aberrantly expressed lncRNAs in MDS and further prioritize biologically relevant lncRNAs using the NLCFA. Taken together, our data suggests that aberrant levels of specific lncRNAs are intimately involved in network modules that control multiple cancer-associated signalling pathways and cellular processes. Importantly, our method can be applied to prioritize aberrantly expressed lncRNAs for functional validation in other diseases and biological contexts.

Availability and implementation: The method is implemented in R language and Matlab.

Contact: xizhou@wakehealth.edu.

Supplementary information: Supplementary data are available at Bioinformatics online.

Fig. 1.

Process for identifying DE lincRNAs in MDS: Left side: Flowchart for identifying DE lincRNAs from RNA-seq datasets. Raw reads were mapped to the human genome (hg19) using Tophat. Expression of lincRNA transcripts was quantified with Cufflinks and DE lincRNAs identified using Cuffdiff. Right side: Process for identifying DE lincRNAs from microarray. All probes in HG-U133Plus2 platform were mapped to lincRNAs using Bowtie2. We kept probes mapped to lincRNAs without mismatch and removed probes which mapped to protein-coding transcripts. LincRNA expression profiles were determined using the re-annotated probes. DE lincRNAs were identified using the limma package in R. DE lincRNAs obtained from RNA-seq and microarrays were combined for further analyses

Fig. 2.

(A) Flowchart for steps in the NLCFA method. DE lincRNAs and miRNAs significantly correlated with protein coding-gene (PCGs) expression were identified. We integrated lincRNAs-PCGs (L-P), miRNAs-PCGs (M-P), pathways-PCGs (P-P) into three different association matrices. These matrices were further analyzed using the MDNMF method to identify co-modules. In each co-module, associated networks were used to annotate lincRNA function. (B) Identified co-modules in MDS. In the three associated matrices, rows correspond to the PCGs and columns correspond to lncRNAs, miRNAs and pathways. These co-module subsets contain highly correlated profiles identified using the MDNMF method. (C) Annotating lincRNA function by a network-based method. For each co-module, we constructed the interaction network and integrated protein-protein interaction. After network propagation, a lincRNAs-GO association score was assigned to each node and each lincRNA was annotated with the prioritized GO terms

Fig. 3.

Chromosomal distribution of DE lincRNAs. The outer circle represents each human chromosome. The blue labels on the outside are the position and name of identified DE lincRNAs. The blue labels on the inside are lincRNAs identified in both RNA-seq and microarray dataset. The inside track is the histogram for the DE lincRNAs on each chromosome. The DE lincRNAs are distributed across all chromosomes. The top three chromosomes that are enriched for DE lincRNAs are chr2 (30 DE lincRNAs), chr1 (22 DE lincRNAs) and chr6 (20 DE lincRNAs)

Fig. 4.

(A) Association network of linc-RPIA identified in co-module 1. In co-module 1, Linc-RPIA is predicted to regulate cancer-associated genes and is itself regulated by miR-16, miR-182 and miR-429. Most associated pathways are associated with the terms cancer or immune system. The blue squares represent pathways, the green triangles represent miRNAs, the yellow circles represent PCGs in that co-module. (B) Connected functional regulatory network for linc-RPIA. The functional networks were merged using Ingenuity Pathway Analysis (IPA). Linc-RPIA is related to TP53 and may also regulate SHC1 via an interaction with miR-429. The rhombus represents linc-RPIA. Connections between miRNAs and PCGs were obtained from IPA (Color version of this figure is available at Bioinformatics online.)