ABSTRACT
ObjectiveInferring cancer driver genes, especially rare or sample-specific cancer driver genes, is crucial for precision oncology. Considering the high inter-tumor heterogeneity, a few recent methods attempt to reveal cancer driver genes at the individual level. However, most of these methods generally integrate multi-omics data into a single biomolecular network (e.g., gene regulatory network or protein-protein interaction network) to identify cancer driver genes, which results in missing important interactions highlighted in different networks. Thus, the development of a multiplex network method is imperative in order to integrate the interactions of different biomolecular networks and facilitate the identification of cancer driver genes. MethodsA multiplex network control method called Personalized cancer Driver Genes with Multiplex biomolecular Networks (PDGMN) was proposed. Firstly, the sample-specific multiplex network, which contains protein-protein interaction layer and gene-gene association layer, was constructed based on gene expression data. Subsequently, somatic mutation data was integrated to weight the nodes in the sample-specific multiplex network. Finally, a weighted minimum vertex cover set identification algorithm was designed to find the optimal set of driver nodes, facilitating the identification of personalized cancer driver genes. ResultsThe results derived from three TCGA cancer datasets indicate that PDGMN outperforms other existing methods in identifying personalized cancer driver genes, and it can effectively identify the rare driver genes in individual patients. Particularly, the experimental results indicate that PDGMN can capture the unique characteristics of different biomolecular networks to improve cancer driver gene identification. ConclusionPDGMN can effectively identify personalized cancer driver genes and broaden our understanding of cancer driver gene identification from a multiplex network perspective. The source code and datasets used in this work are available at https://github.com/NWPU-903PR/PDGMN.
ABSTRACT
OBJECTIVE@#To study the expression and function of long non-coding RNA linc00467 in childhood acute myeloid leukemia (AML).@*METHODS@#Bone marrow samples were collected from 5 children with AML who were diagnosed from May 2016 to June 2018. Normal bone marrow samples based on bone marrow examination were collected from 3 children as controls. Quantitative real-time PCR was used to measure the expression of linc00467 in the two groups. A lentivirus system was used to achieve overexpression of linc00467 in AML cells (HL-60) (linc00467 overexpression group), and empty vector expressing green fluorescent protein (GFP) was transfected into AML cells to establish a GFP control group. A lentivirus system was used to insert an interfering sequence into AML cells (sh-linc00467 interfering group), and a random sequence was inserted to establish an sh-NC control group. Cell proliferation and resistance to doxorubicin were observed for all groups.@*RESULTS@#Compared with the normal control group, the children with AML had a significant increase in linc00467 (P=0.018). Overexpression and interference with linc00467 expression had no significant effect on cell proliferation. Compared with the GFP control group, the linc00467 overexpression group had a significant increase in the viability of HL-60 cells at the adriamycin concentrations of 0.1, 0.2, 0.3, 0.4, and 0.5 μg/mL (P<0.05). Compared with the sh-NC control group, the sh-linc00467 interfering group had a significant reduction in the viability of HL-60 cells at the adriamycin concentrations of 0.1, 0.2, 0.3, 0.4, and 0.5 μg/mL (P<0.05). Compared with the untreated group, the adriamycin treatment group had a significant increase in the expression of linc00467 in HL-60 cells (P<0.05).@*CONCLUSIONS@#This study reveals the biological function of linc00467 to promote the resistance to adriamycin in AML, which provides a basis for developing new therapeutic drugs for AML.