By collecting multiple sets per subject in microarray data, gene sets analysis requires characterize intra-subject variation using gene expression profiling. For each subject, the data can be written as a matrix with the different subsets of gene expressions (e.g. multiple tumor types) indexing the rows and the genes indexing the columns. To test the assumption of intra-subject (tumor) variation, we present and perform tests of multi-set sphericity and multi-set identity of covariance structures across subjects (tumor types). We demonstrate by both theoretical and empirical studies that the tests have good properties. We applied the proposed tests on The Cancer Genome Atlas (TCGA) and tested covariance structures for the gene expressions across several tumor types.
BarryWNobelAWrightF.Significance analysis of functional categories in gene expression studies: a structured permutation approach. Bioinformatics2005;
21: 1943–1949.
2.
NewtonMQuintanaFDen BoonJ, et al.Random-set methods identify distinct aspects of the enrichment signal in gene-set analysis. Ann Appl Stat2007;
1: 85–106.
3.
EfronBTibshiraniR.On testing the significance of sets of genes. Ann Appl Stat2007;
1: 107–129.
4.
DasSMcClainCJRaiSN.Fifteen years of gene set analysis for high-throughput genomic data: a review of statistical approaches and future challenges. Entropy2020;
22: 427.
5.
MalekiFOvensKHoganDJ, et al.Gene set analysis: challenges, opportunities, and future research. Frontier Genet2020;
11: 1–16.
6.
MenasheIMaederDGarcia-ClosasM, et al.Pathway analysis of breast cancer genome-wide association study highlights three pathways and one canonical signaling cascade. Cancer Res2010; 70: 4453–4459.
7.
SullivanPFPosthumaD.Biological pathways and networks implicated in psychiatric disorders. Curr Opin Behav Sci2015;
2: 58–68.
8.
de LeeuwCANealeBMHeskesT, et al.The statistical properties of gene-set analysis. Nat Rev Genet2016;
17: 353–364.
9.
TouloumisATavar´eSMarioniJC.Testing the mean matrix in high-dimensional transposable data. Biometrics2015;
71: 157–166.
VogelsteinBPapadopoulosKVelculescuVE, et al.Cancer genome landscapes. Science2013;
339: 1546–1558.
12.
WheelerDAWangL.From human genome to cancer genome: The first decade. Genome Res2013;
23: 1054–1062.
13.
ChangKCreightonCDavisC, et al.
The cancer genome atlas pan-cancer analysis project. Nat Genet2013;
45: 1113–1120.
14.
de Torrent, L, Zimmerman S, Suzuki M, et al. The shape of gene expression distributions matter: how incorporating distribution shape improves the interpretation of cancer transcriptomic data. BMC Bioinform 2020; 21: 562.
15.
BirneyEClampMDurbinR.GeneWise and genomewise. Genome Res2004;
14: 988–995.
16.
ChenSXZhangLXZhangPS.Tests for high-dimensional covariance matrices. J Am Stat Assoc2010;
105: 810–819.
17.
QiuYChenSX.Test for bandedness of high-dimensional covariance matrices and bandwidth estimation. Ann Stat2012;
40: 1285–1314.
18.
LedoitOWolfM.Some hypothesis tests for the covariance matrix when the dimension is large compared to the sample size. Ann Stat2002;
30: 1081–1102.
19.
SrivastavaMS.Some tests concerning the covariance matrix in high dimensional data. J Japan Statist Soc2005;
35: 251–272.
20.
CaiTLiuWXiaY.Two-sample covariance matrix testing and support recovery in high-dimensional and sparse settings. J Am Stat Assoc2013;
108: 265–277.
21.
CaiTMaZM.Optimal hypothesis testing for high dimensional covariance matrices. Bernoulli2013;
19: 2359–2388
IshllAYataKAoshimaM.Equality tests of high-dimensional covariance matrices under the strongly spiked eigenvalue model. J Stat Plan Inference2019;
202: 99–111.
24.
ZhangQYHuJBaiZD. Modified Pillai’s trace statistics for two high-dimensional sample covariance matrices, arXiv:2001.00397v1 [math.ST] 2020.
25.
DonohoDJinJ.Higher criticism for detecting sparse heterogeneous mixtures. Ann Stat2004;
32: 962–994.
26.
ZhongPSChenSXXuM.Tests alternative to higher criticism for high-dimensional means under sparsity and column-wise dependence. Ann Stat2013;
6: 2820–2851.
27.
AhmadMR.Testing homogeneity of several covariance matrices and multi-sample sphericity for high-dimensional data under non-normality. Comm Stat Theory Meth2017;
46: 3738–3753.
28.
BoxGEP.A general distribution theory for a class of likelihood criteria. Biometrika1949;
36: 317–346.
29.
BaiZJiangDYaoY, et alCorrections to LRT on large-dimensional covariance matrix by RMT. Ann Stat2009; 37: 3822–3840.
30.
JiangDJiangTYangF.Likelihood ratio tests for covariance matrices of high-dimensional normal distributions. J Stat Plan Inference2012;
142: 2241–2256.
31.
JiangTYangF.Central limit theorems for classical likelihood ratio tests for high-dimensional normal distributions. Ann Stat2013;
41: 2029–2074.
32.
BerkesILiuWWuWB.Major-Tusnády approximation under dependence. Ann Probabil2014;
42: 794–817.
33.
LiuWLinZ.Strong approximation for a class of stationary processes. Stochastic Process Appl2009;
119: 249–280.
34.
DykstraRL.Establishing the positive definiteness of the sample covariance matrix. Ann Math Statist1970;
41: 2153–2154.
35.
QayedAHanD. Homogeneity test of several high-dimensional covariance matrices for stationary processes under non-normality 2020. arXiv:2008.09259 [math.ST]
36.
GuptaAKNagarDK.Matrix variate distribution.
New York, NY:
Chapman and Hall/CRC, 2000.
37.
PetrochilosDShojaieAGennariJ, et al.Using random walks to identify cancer-associated modules in expression data. BioData Mining2013;
6, https://doi.org/10.1186/1756-0381-6-17
38.
DongFHeYWangT, et al.
Predicting viral exposure response from modeling the changes of co-expression networks using time series gene expression data. BMC Bioinform2020;
21: 370.
39.
BartlettMS.Properties of sufficiency and statistical tests. Proc Royal Soc Lond Series A1937;
160: 268–282.
40.
AhmadMR.A U-statistic approach for a high-dimensional two-sample mean testing problem under non-normality and Behrens-Fisher setting. Ann Inst Stat Math2014;
66: 33–61.
41.
HuberWHeydebreckASültmannH, et al.Variance stabilization applied to microarray data calibration and to the quantification of differential expression. Bioinformatics2002;
18: S96–S104.
42.
BolstadBMIrizarryRAAstrandM, et al.A comparison of normalization methods for high density oligonucleotide array data based on bias and variance. Bioinformatics2003;
19: 185–193.
43.
FarahbodMPavlidisP.Differential coexpression in human tissues and the confounding effect of mean expression levels. Bioinformatics2019;
35: 55–61, https://doi.org/10.1093/bioinformatics/bty538
44.
Chao M and Glaser R. The exact distribution of Bartlett’s test statistic for homogeneity of variances with unequal sample sizes. J Am Stat Assoc 1978; 73: 422–426.
