Integrated miRNA–risk gene–pathway pair network analysis provides prognostic biomarkers for gastric cancer

Figures & data

Table 1 Summary of mRNA expression profiles for gastric cancer

ID	Number
	Disease sample	Control sample	Total
GSE51575	26	26	52
GSE65801	32	32	64
GSE33651	40	12	52
GSE63089	45	45	90
GSE56807	5	5	10
GSE33335	25	25	50
GSE27342	80	80	160
GSE13195	25	25	50
GSE30727	30	30	60
GSE19826	12	15	27
GSE13911	38	31	69
GSE38932	24	24	48
GSE37023	141	75	216

Table 2 Top ten differentially expressed genes

Gene	FDR
COL4A1	6.65E-26
SERPINH1	1.23E-26
IFITM3	4.07E-25
LY6E	4.81E-23
COL1A1	1.37E-23
COL1A2	1.37E-23
SULF1	3.98E-21
ATP4B	2.16E-21
CTSB	2.65E-20
FN1	1.64E-20

Abbreviation: FDR, false discovery rate.

Table 3 Top ten differentially expressed miRNAs

miRNA	FDR
hsa-miR-204	1.02E-13
hsa-miR-1246	8.43E-12
hsa-miR-135b	3.35E-07
hsa-miR-30c	5.94E-07
hsa-miR-29c	5.94E-07
hsa-miR-646	7.43E-07
hsa-miR-148a	8.09E-07
hsa-miR-363	2.22E-06
hsa-miR-1206	5.14E-06
hsa-miR-196b	6.58E-06

Abbreviation: FDR, false discovery rate.

Figure 1 The numbers of target genes for DEmiRs.

Notes: In all, (A) a total of 8,642 predicted target genes for DEmiRs were identified from PicTar, PITA, miRanda, and TargetScan, while (B) a total of 912 verified target genes were identified from miRecords, miRTarBase, and Tarbase. Finally, (C) the common genes of potential targets and verified targets were selected for the eventual miRNA–target pairs.
Abbreviation: DEmiR, differentially expressed miRNA.

Figure 2 The number of risk genes.

Abbreviations: DEG, differentially expressed gene; DEmiR, differentially expressed microRNA.

Figure 3 Enriched functions and pathways of risk genes.

Notes: (A) Biological process; (B) cellular components; (C) molecular functions; (D) pathways. The terms are listed in the Y-axis, while the significances are listed on the X-axis. The higher the log(P-value), the more significant is the term.
Abbreviation: ECM, extracellular matrix.

Figure 4 Heat map of the pathway expression profile.

Notes: The X-axis refers to the pathways, and the Y-axis refers to the samples. The colors from green to red indicate the pathway expression level from low to high.
Abbreviations: GABA, gamma-aminobutyric acid; TGF-β, transforming growth factor-β; VEGF, vascular endothelial growth factor; cAMP, 3′,5′-cyclic adenosine monophosphate; TNF, tumor necrosis factors; GnRH, gonadotropin-releasing hormone; TRP channels, transient receptor potential channels; AMPK, AMP-activated protein kinase; NOD, nucleotide-binding, oligomerization domain; RIG-I, retinoic acid-inducible gene 1; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; ECM, extracellular matrix; PI3K, phosphatidylinositol 3-kinase; ABC transporters, ATP-binding cassette transporters; HTLV-I, human T-cell lymphotropic virus type 1; cGMP, 3′,5′-cyclic guanosine monophosphate; PPAR, peroxisome proliferator-activated receptors; Jak-STAT, janus kinase-signal transducer and activator of transcription; MAPK: mitogen-activated protein kinase; mTOR, mammalian target of rapamycin.

Figure 5 miRNA-risk gene-pathway pair complex network.

Notes: The rhombus refers to the miRNA, the rotundity refers to the mRNA, and yellow refers to the co-pathway pairs. Red indicates upregulation and green indicates downregulation.
Abbreviation: ECM, extracellular matrix.

Figure 6 Interactive nodes of risk gene CALM2 in the integrated network of miRNA–risk gene–pathway pair.

Notes: The rhombus refers to the miRNA, the rotundity refers to the mRNA, and yellow refers to the co-pathway pairs. Red indicates upregulation and green indicates downregulation.
Abbreviation: ECM, extracellular matrix.

Figure 7 Expressions and classification efficacy of two miRNAs with highest degrees in the integrated network.

Notes: (A) Expression levels between gastric cancer samples and normal samples; (B) ROC curves of the miRNAs for the gastric cancer samples.
Abbreviation: ROC, receiver operating characteristic.

Table 4 Network topological features of 15 pathways

Pathway pair	Average shortest path length	Betweenness centrality	Closeness centrality	Degree	Topological coefficient
Focal adhesion/ECM–receptor interaction	3.793734	0.026451	0.263593	11	0.142857
Proteoglycans in cancer/microRNAs in cancer	3.903394	0.012734	0.256187	7	0.18254
cGMP–PKG signaling pathway/cAMP signaling pathway	3.767624	0.008914	0.265419	6	0.233333
Ras signaling pathway/Rap1 signaling pathway	3.762402	0.010502	0.265788	6	0.219298
Cell cycle/oocyte meiosis	4.331593	0.006181	0.230862	5	0.225
Estrogen signaling pathway/melanogenesis	3.929504	0.004038	0.254485	4	0.352941
Cholinergic synapse/serotonergic synapse	4.503916	0.001182	0.222029	3	0.484848
Adherens junction/tight junction	4.691906	0.001816	0.213133	2	0.5
Chronic myeloid leukemia/acute myeloid leukemia	4.561358	0.000973	0.219233	2	0.5
Dopaminergic synapse/long-term depression	4.775457	0.000139	0.209404	2	0.8
MAPK signaling pathway/ErbB signaling pathway	4.723238	0.00051	0.211719	2	0.5
NF-kappa B signaling pathway/HIF-1 signaling pathway	4.921671	0.000103	0.203183	2	0.666667
Nucleotide excision repair/mismatch repair	5.929504	6.83E-06	0.168648	2	1
Oocyte meiosis/p53 signaling pathway	4.759791	0.000231	0.210093	2	0.666667
Pancreatic cancer/endometrial cancer	4.895561	0	0.204267	1	0

Abbreviations: cAMP, 3′,5′-cyclic adenosine monophosphate; cGMP, 3′,5′-cyclic guanosine monophosphate; ECM, extracellular matrix; PKG, protein kinase G; MAPK, mitogen-activated protein kinase; HIF-1, hypoxia-inducible factor-1.

Table 5 Network topological features of the top 15 risk genes with highest degrees

Risk_gene	Average shortest path length	Betweenness centrality	Closeness centrality	Degree	Topological coefficient
CALM2	3.328982	0.043302	0.300392	9	0.128042
MAP2K1	3.898172	0.022464	0.25653	9	0.133903
NSMAF	3.402089	0.027349	0.293937	7	0.202381
COL1A2	3.694517	0.016129	0.270671	6	0.236111
FKBP1A	3.365535	0.034681	0.29713	6	0.20405
SOX4	3.240209	0.044469	0.308622	6	0.206037
CCND2	3.245431	0.028494	0.308126	5	0.233871
GATA6	3.537859	0.019326	0.282657	5	0.220513
GNAI2	3.887728	0.008331	0.25722	5	0.232432
GNAI3	4.221932	0.005805	0.236858	5	0.257143
ITGA5	4.007833	0.011827	0.249511	5	0.276923
KPNA3	3.292428	0.023322	0.303727	5	0.25812
RND3	3.590078	0.036432	0.278545	5	0.221739
THBS1	3.386423	0.03582	0.295297	5	0.212121
WEE1	3.240209	0.040427	0.308622	5	0.2304

Table 6 Network topological features of the top 15 differentially expressed miRNAs with highest degrees

DEmiR	Average shortest path length	Betweenness centrality	Closeness centrality	Degree	Topological coefficient
hsa-miR-19b	2.906005	0.188733	0.344115	47	0.049156
hsa-miR-181b	2.963446	0.184998	0.337445	44	0.054075
hsa-miR-20b	2.942559	0.174251	0.33984	44	0.067149
hsa-miR-30e	3.036554	0.120754	0.329321	33	0.055944
hsa-miR-421	3.214099	0.120532	0.311129	32	0.0625
hsa-miR-141	3.308094	0.125676	0.302289	31	0.052227
hsa-miR-101	3.229765	0.106094	0.30962	30	0.063333
hsa-miR-497	3.45953	0.092437	0.289057	25	0.052727
hsa-miR-204	3.234987	0.069288	0.30912	20	0.07619
hsa-miR-92b	3.32376	0.047964	0.300864	19	0.105263
hsa-miR-135a	3.396867	0.036556	0.294389	19	0.134503
hsa-miR-224	3.438642	0.050803	0.290812	18	0.104575
hsa-miR-133a	3.605744	0.056861	0.277335	17	0.079585
hsa-miR-363	3.412533	0.032906	0.293037	16	0.125
hsa-miR-339-5p	3.94517	0.061805	0.253475	15	0.081481

Abbreviation: DEmiR, differentially expressed miRNA.