Massive amounts of high-dimensional data have been accumulated over the past two decades, which has cultured increasing interests in identifying gene pathways related to certain biological processes. In particular, since pathway-based analysis has the ability to detect subtle changes of differentially expressed genes that could be missed when using gene-based analysis, detecting the gene pathways that regulate certain diseases can provide new strategies for medical procedures and new targets for drug discovery. Limited work has been carried out, primarily in regression settings, to study the effects of pathways on survival outcomes. Motivated by a breast cancer gene-pathway data set, which exhibits the “small n, large p” characteristics, we propose a semiparametric Bayesian kernel survival model (s-BKSurv) to study the effects of both clinical covariates and gene expression levels within a pathway on survival time. We model the unknown high-dimensional functions of pathways via Gaussian kernel machine to consider the possibility that genes within the same pathway interact with each other. To address the multiple comparisons problem under a full Bayesian setting, we propose a similarity-dependent procedure based on Bayes factor to control the family-wise error rate. We demonstrate the outperformance of our approach under various simulation settings and pathways data.
EfronB. The efficiency of cox's likelihood function for censored data. J Am Stat Assoc1977; 72: 557–565.
2.
TibshiraniR. Regression shrinkage and selection via the lasso. J R Stat Soc Series B Stat Methodol1996; 58: 267–288. .
3.
FanJLiR. Variable selection via nonconcave penalized likelihood and its oracle properties. J Am Stat Assoc2001; 96: 1348–1360.
4.
ZouHHastieT. Regularization and variable selection via the elastic net. J R Stat Soc Series B Stat Methodol2005; 67: 301–320.
5.
GuiJAndrewASAndrewsP, et al.A robust multifactor dimensionality reduction method for detecting gene–gene interactions with application to the genetic analysis of bladder cancer susceptibility. Ann Hum Genet2011; 75: 20–28.
6.
LeeSKwonMSOhJM, et al.Gene–gene interaction analysis for the survival phenotype based on the cox model. Bioinformatics2012; 28: i582–i588.
LiuKQLiuZPHaoJK, et al.Identifying dysregulated pathways in cancers from pathway interaction networks. BMC Bioinf2012; 13: 126.
9.
WangJZuoYManYG, et al.Pathway and network approaches for identification of cancer signature markers from omics data. J Cancer2015; 6: 54.
10.
GoemanJJVan De GeerSADe KortF, et al.A global test for groups of genes: testing association with a clinical outcome. Bioinformatics2004; 20: 93–99.
11.
SubramanianATamayoPMoothaVK, et al.Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A2005; 102: 15545–15550.
12.
PangHLinAHolfordM, et al.Pathway analysis using random forests classification and regression. Bioinformatics2006; 22: 2028–2036.
13.
LiuDLinXGhoshD. Semiparametric regression of multidimensional genetic pathway data: Least-squares kernel machines and linear mixed models. Biometrics2007; 63: 1079–1088.
14.
TaiFPanW. Incorporating prior knowledge of predictors into penalized classifiers with multiple penalty terms. Bioinformatics2007; 23: 1775–1782.
15.
KempDMNirmalaNSzustakowskiJD. Extending the pathway analysis framework with a test for transcriptional variance implicates novel pathway modulation during myogenic differentiation. Bioinformatics2007; 23: 1356–1362.
16.
WeiZLiH. A markov random field model for network-based analysis of genomic data. Bioinformatics2007; 23: 1537–1544.
17.
WuMCZhangLWangZ, et al.Sparse linear discriminant analysis for simultaneous testing for the significance of a gene set/pathway and gene selection. Bioinformatics2009; 25: 1145–1151.
18.
StingoFCChenYATadesseMG, et al.Incorporating biological information into linear models: a Bayesian approach to the selection of pathways and genes. Ann Appl Stat2011, pp. 5. doi: 0.1214/11-AOAS463.
19.
KimIPangHZhaoH. Bayesian semiparametric regression models for evaluating pathway effects on continuous and binary clinical outcomes. Stat Med2012; 31: 1633–1651.
20.
KimIPangHZhaoH. Statistical properties on semiparametric regression for evaluating pathway effects. J Stat Plan Inference2013; 143: 745–763.
21.
PangHKimIZhaoH. Random effects model for multiple pathway analysis with applications to type ii diabetes microarray data. Stat Biosci2015; 7: 167–186.
22.
ChengLKimIPangH. Bayesian semiparametric model for pathway-based analysis with zero-inflated clinical outcomes. J Agric Biol Environ Stat2016; 21: 641–662.
23.
GoemanJJOostingJCleton-JansenAM, et al.Testing association of a pathway with survival using gene expression data. Bioinformatics2005; 21: 1950–1957.
24.
PangHDattaDZhaoH. Pathway analysis using random forests with bivariate node-split for survival outcomes. Bioinformatics2010; 26: 250–258.
25.
CaiTToniniGLinX. Kernel machine approach to testing the significance of multiple genetic markers for risk prediction. Biometrics2011; 67: 975–986.
26.
LinXCaiTWuMC, et al.Kernel machine SNP-set analysis for censored survival outcomes in genome-wide association studies. Genet Epidemiol2011; 35: 620–631.
27.
MercerJ. Functions of positive and negative type, and their connection with the theory of integral equations. Philos Trans R Soc London, Ser A1909; 209: 415–446.
28.
BickelPJLevinaE. Covariance regularization by thresholding. Ann Stat2008; 36: 2577–2604. .
29.
FriedmanJHastieTTibshiraniR. The elements of statistical learning2001; Vol 1, Berlin: Springer series in statistics Springer.
30.
KassRERafteryAE. Bayes factors. J Am Stat Assoc1995; 90: 773–795.
31.
NewtonMARafteryAE. Approximate Bayesian inference with the weighted likelihood bootstrap. J R Stat Soc Series B Stat Methodol1994; 56: 3–48. .
32.
HowladerNNooneAKrapchoM, et al.Seer cancer statistics review, 1975–2015, National Cancer Institute, Bethesda, MD: National Cancer Institute, 2015.
33.
LiGSunY. A simulation-based goodness-of-fit test for survival data. Stat Probab Lett2000; 47: 403–410.
34.
LeméeFBergoglioVFernandez-VidalA, et al.Dna polymerase θ up-regulation is associated with poor survival in breast cancer, perturbs DNA replication, and promotes genetic instability. Proc Natl Acad Sci U S A2010; 107: 13390–13395.
35.
LangeSSTakataKiWoodRD. DNA polymerases and cancer. Nat Rev Cancer2011; 11: 96–110.
36.
LinJLeeIMSongY, et al.Plasma homocysteine and cysteine and risk of breast cancer in women. Cancer Res2010; 70: 2397–2405. .
37.
JeschkeJO'HaganHMZhangW, et al.Frequent inactivation of cysteine dioxygenase type 1 contributes to survival of breast cancer cells and resistance to anthracyclines. Clin Cancer Res2013; 19: 3201–3211.
38.
BraitMLingSNagpalJK, et al.Cysteine dioxygenase 1 is a tumor suppressor gene silenced by promoter methylation in multiple human cancers. PloS One2012; 7: e44951.
39.
Zamora-LeónSPGoldeDWConchaII, et al.Expression of the fructose transporter glut5 in human breast cancer. Proc Natl Acad Sci U S A1996; 93: 1847–1852.
40.
LiuHHuangDMcArthurDL, et al.Fructose induces transketolase flux to promote pancreatic cancer growth. Cancer Res2010; 70: 6368–6376.
41.
YagitaMNakajimaMSakselaE. Suppression of human natural killer cell activity by amino sugars. Cell Immunol1989; 122: 83–95.
42.
MamessierESylvainAThibultML, et al.Human breast cancer cells enhance self tolerance by promoting evasion from NK cell antitumor immunity. J Clin Invest2011; 121: 3609–3622.
43.
TroupSNjueCKliewerEV, et al.Reduced expression of the small leucine-rich proteoglycans, lumican, and decorin is associated with poor outcome in node-negative invasive breast cancer. Clin Cancer Res2003; 9: 207–214.
44.
CarlsonECLiuCYChikamaT, et al.Keratocan, a cornea-specific keratan sulfate proteoglycan, is regulated by lumican. J Biol Chem2005; 280: 25541–25547.
45.
ChoiHJChungTWKimCH, et al.Estrogen induced β-1, 4-galactosyltransferase 1 expression regulates proliferation of human breast cancer mcf-7 cells. Biochem Biophys Res Commun2012; 426: 620–625.
46.
Diarra-MehrpourMArrabalSJalilA, et al.Prion protein prevents human breast carcinoma cell line from tumor necrosis factor α-induced cell death. Cancer Res2004; 64: 719–727.
47.
JollietPSimonNBarreJ, et al.Plasma coenzyme q10 concentrations in breast cancer: prognosis and therapeutic consequences. Int J Clin Pharmacol Ther1998; 36: 506–509.
48.
LopesNSousaBMartinsD, et al.Alterations in vitamin d signalling and metabolic pathways in breast cancer progression: a study of vdr, cyp27b1 and cyp24a1 expression in benign and malignant breast lesions vitamin d pathways unbalanced in breast lesions. BMC Cancer2010; 10: 483.
49.
DeebKKTrumpDLJohnsonCS. Vitamin d signalling pathways in cancer: potential for anticancer therapeutics. Nat Rev Cancer2007; 7: 684–700.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
