You are viewing a javascript disabled version of the site. Please enable Javascript for this site to function properly.
Go to headerGo to navigationGo to searchGo to contentsGo to footer
In content section. Select this link to jump to navigation

Cancer classification and biomarker selection via a penalized logsum network-based logistic regression model

Article type: Research Article

Authors: Zhou, Zhiminga | Huang, Haihuia; b | Liang, Yongc; *

Affiliations: [a] Faculty of Information Technology, Macau University of Science and Technology, Macau, China | [b] Shaoguan University, Shaoguan, Guangdong, China | [c] Macau Institute of Systems Engineering and Collaborative Laboratory of Intelligent Science and Systems, Macau University of Science and Technology, Macau, China

Correspondence: [*] Corresponding author: Yong Liang, Macau Institute of Systems Engineering and Collaborative Laboratory of Intelligent Science and Systems, Macau University of Science and Technology, Macau, China. E-mail: [email protected].

Keywords: Regularization, gene selection, log-sum penalty, network-based knowledge

DOI: 10.3233/THC-218026

Journal: Technology and Health Care, vol. 29, no. S1, pp. 287-295, 2021

Published: 25 March 2021
Get PDF open access

Abstract

BACKGROUND:

In genome research, it is particularly important to identify molecular biomarkers or signaling pathways related to phenotypes. Logistic regression model is a powerful discrimination method that can offer a clear statistical explanation and obtain the classification probability of classification label information. However, it is unable to fulfill biomarker selection.

OBJECTIVE:

The aim of this paper is to give the model efficient gene selection capability.

METHODS:

In this paper, we propose a new penalized logsum network-based regularization logistic regression model for gene selection and cancer classification.

RESULTS:

Experimental results on simulated data sets show that our method is effective in the analysis of high-dimensional data. For a large data set, the proposed method has achieved 89.66% (training) and 90.02% (testing) AUC performances, which are, on average, 5.17% (training) and 4.49% (testing) better than mainstream methods.

CONCLUSIONS:

The proposed method can be considered a promising tool for gene selection and cancer classification of high-dimensional biological data.

1.Introduction

Microarray technology is one of the most recent advances in cancer research, and using this method, the expression levels of thousands of genes can be recorded simultaneously. In genomic analysis, the identification of molecular biomarkers or signal pathways associated with phenotypes is a particularly important issue. Logistic regression is a powerful discrimination method, enables clear statistical interpretation, and derives classification probability of classification label information.

From a biological point of view, only a few genes are related to the target disease, and most genes are not involved in cancer classification. Unrelated genes may cause noise and reduce the accuracy of prediction. In addition, from a machine learning perspective, too many features may cause overfitting and negatively affect classification performance.

The regularization method has been widely used when dealing with high-dimensional problems. A popular regularization method is the absolute shrinkage and selection operator (Lasso) or L1 penalty [1], which can simultaneously perform feature selection and model construction. L1 penalty extensions, such as the SCAD penalty [2], which is symmetrical, non-concave, have also been common for many years. The adaptive Lasso [3] uses dynamic weights to penalize different coefficients. However, in some cases, L1 type regularization may result in inconsistent feature selection and often introduce extra bias in parameter estimation into the statistical model [4]. Xu et al. [4] suggested L1/2 regularization, which and can produce a more accurate solution than the L1 penalty [5, 6, 7]. But in the analysis of gene expression data, L1/2 regularization may not be sparse enough. In theory, L0 regularization produces more sparse and better solutions [8], but this is an NP problem. Therefore, Candes et al. suggested the logsum penalty [9], a method that nicely approximates the L0 penalty. The logsum penalty has been successfully applied in much research in recent years, such as impact force recognition [10], drug discovery [11], etc. However, the logsum penalty lacks a mechanism to incorporate prior knowledge of genetic biology. Combining gene expression with the analysis of network knowledge can reduce noise and detect complicating factors in genomic data analysis of many regression and classification models [12, 13, 14, 15].

In short, existing methods show good results in terms of feature selection and model construction, but they either cannot produce sufficient sparsity or do not use any network interaction knowledge.

In this paper, we investigate the sparse logistic regression model with a logsum network-based (Logsum-Net) penalty, in particular for gene selection in cancer classification.

The major contributions of this paper are as follows:

  • 1. A new method of gene selection is put forward. The logsum method is an efficient tool for feature selection; however, this method is recommended from a strictly computational view, and there is no built-in design that can use a priori biological structure information. Unlike previous research, in this research, a Logsum-Net regularization that aims to integrate prior biological graph information is proposed.

  • 2. Beyond several mainstream methods, we have carried out a simulation experiment and experimented on a breast gene expression data set, and the experimental results show that the method is feasible. For the breast cancer data set, the AUC of our method is, on average, 5.17% (training) and 4.49% (testing) better than mainstream methods.

2.The penalized logsum network-based logistic regression model

Suppose that dataset D has n samples D ={(X1,y1),(X2,y2),…,(Xn,yn)}, where Xi=(xi⁢1,xi⁢2,…,xi⁢p) is the ith sample with p genes and yi is the corresponding dependent variable consisting of a binary value of either zero or 1. Define a classifier f⁢(x)=ex/(1+ex), and the logistic regression is defined as:

(1)
P(yi=1|Xi)=f(Xi′β)=exp⁡(Xi′⁢β)1+exp⁡(Xi′⁢β)

where β=(β1,…,βp) are the estimated variables. We transform Eq. (1) using simple algebra:

(2)
l⁢(β)=-∑i=1n{yi⁢log⁡[f⁢(Xi′⁢β)]+(1-yi)⁢log⁡[1-f⁢(Xi′⁢β)]}

Equation (2) is easily to overfitted when applied to a problem with high dimensions and low sample size. Regularization is a commonly used technique to solve high-dimensional problems, which can be expressed as:

(3)
L⁢(λ,β)=l⁢(β)+P⁢(β),

where P⁢(β) is the regularization or penalty term. A typical penalty is the Lasso (L1) method [1] which has the form Pλ,𝐿𝑎𝑠𝑠𝑜⁢(β)=λ⁢∑j=1p|βj|1, where λ> 0. With the singularity characteristic of L1 regularization, an L1 penalized regression model can force small coefficients to zero. There are various versions of the L1 penalty, such as elastic net [16], SCAD [2]. However, L1 type regularization has two drawbacks: it is biased and may not be sparse enough. Xu et al. [4] proposed an L1/2 method of the form Pλ,L⁢1/2⁢(β)=λ⁢∑j=1p|βj|1/2, which is able to generate more a sparser solution than L1 methods. However, in the analysis of genomics data, L1/2 regularization may not be sparse enough. Theoretically, L0 regularization produces better solutions with more sparsity, but this is an NP problem. Therefore, Candes et al. [9] proposed the logsum penalty, which approximates L0 regularization much better. The logsum regularization is shown as follows:

(4)
Pλ,𝐿𝑜𝑔𝑠𝑢𝑚⁢(β)=λ⁢∑j=1plog⁡(|βj|+ε),

where ε> 0 should be set arbitrarily small to make the logsum penalty closely resemble the L0-norm. However, the logsum penalty is unable to use any prior biological knowledge such as the gene-regulatory network, as this method was proposed from a purely computational point of view.

There has been much research into network-based penalties. Li and Li [17], Chen et al. [18] and Wang et al. [19], for example, suggested a Lasso network-based method for the study of gene expression. However, the result achieved by the Lasso method is not sparse enough for genome data. In this paper, we propose a logsum network-based (Logsum-Net) method as follows:

(5)
Pλ1,λ2,Logsum-net⁢(β)=Pλ1,𝐿𝑜𝑔𝑠𝑢𝑚⁢(β)+λ2⁢β⁢L⁢β,

where L is a symmetric Laplace matrix that combines the knowledge of biological network, and the β⁢L⁢β term forces a smooth result on the network.

(6)
Pλ1,λ2,Logsum-net⁢(β)=Pλ1,𝐿𝑜𝑔𝑠𝑢𝑚⁢(β)+λ2⁢∑1⩽i<k⩽p;wi⁢k⁢(βidi-βkdk)2,

where wi⁢k∈[0,1] depends on whether features i and k are linked or not; di and dk denote the degree (includes out or in degree) of features i and k; λ1 and λ2 are penalty tuning parameters. The logsum network-based logistic regression model (Logsum-NL) can be formed as follows:

(7)
β^=argminβ⁢{l⁢(β)+Pλ1,λ2,Logsum-Net⁢(β)},

This equation not only ensures sparsity in the solutions and, making them more appropriate for biological interpretation, but also smooths the regression coefficient of the genes that are connected in the network.

3.Algorithm

We first proposed a new threshold-based solver of the Logsum-Net penalty. Then, we applied an efficient coordinate descent algorithm (CDA) [20] to solve the Logsum-NL.

3.1Threshold solver of the Logsum-Net penalty

Assuming a linear model with p predictors:

y=x1⁢β1+x2⁢β2+…+xp⁢βp,

For simplicity, the predictors and responses are all standardized. A Logsum-Net linear model can be expressed as:

(8)
ℒ⁢(β)=12⁢||ω^-β||2+Pλ1,λ2,Logsum-Net⁢(β),

where y^=X⁢ω^ and ω^=XT⁢y.

Recall that Pλ1,λ2,Logsum-Net⁢(β)=Pλ1,𝐿𝑜𝑔𝑠𝑢𝑚⁢(β)+λ2⁢∑1⩽i<k⩽p;wi⁢k⁢(βidi-βkdk)2, which can be rewritten as:

Pλ1,λ2,Logsum-Net⁢(β)=Pλ1,𝐿𝑜𝑔𝑠𝑢𝑚⁢(β)+λ2⁢∑i=1pwi⁢j⁢(βidi-βjdj)2
(9)
 +λ2⁢∑1⩽i<k⩽p;i,k≠jwi⁢k⁢(βidi-βkdk)2.

The first partial derivative concerning βj of Eq. (8) is given by:

(10)
∂∂⁡βj⁢ℒ⁢(β)=βj-ω^+λ1⁢1βj+ε+2⁢λ2⁢βj-t,

where t=λ2⁢∑i=1pwi⁢j⁢βidi⁢dj.

By setting Eq. (3.1) = 0, a threshold-based solver for the jth item in Logsum-Net penalized linear regression model can be shown as follows:

(11)
βj={𝑠𝑖𝑔𝑛⁢(ω^j)⁢c1+c22if ⁢c2>00if ⁢c2⩽0,

where c1=ω^+t-ε-2⁢λ2⁢ε1+2⁢λ2, c2=c12-4⁢(λ1-ω^j⁢ε-t⁢ζ1+2⁢λ2).

3.2Algorithm for the Logsum-Net penalized logistic regression model

By the Taylor series method, we can rewrite Eq. (2) as follows:

(12)
l⁢(β)≈12⁢n⁢∑i=1n(Zi-Xi⁢β)′⁢Wi⁢(Zi-Xi⁢β)

where Zi=Xi⁢β~+yi-f⁢(Xi⁢β~)f(Xiβ~)(1-f(Xiβ~) is the estimated response, and Wi=f(Xiβ~)(1-f(Xiβ~) is the weight for the estimated response. f⁢(Xi⁢β~)=exp⁡(Xi⁢β~)/(1+exp⁡(Xi⁢β~)) is the evaluated value under the current parameters. Thus, we can redefine the partial residual for fitting the current β~ as Zˇi(j)=∑k≠jxi⁢k⁢β~k and ωj=∑i=1nWi⁢xi⁢j⁢(Zi-Zˇi(j)). The whole algorithm for the Logsum-NL as follows:

Step 1: Set all βj←0 (j=1,2,…,p) and X,y,ε. m←0, λ1 and λ2 are chosen by gird search; Step 2: Calculate Z⁢(m) and W⁢(m) based on the current β⁢(m); Step 3: Update each βj⁢(m) and cycle over j=1,…,p Step 3.1: Calculate Zˇi(j)⁢(m)←∑k≠jxi⁢k⁢βk⁢(m), t=λ2⁢∑i=1pwi⁢j⁢βidi⁢dj. ωj⁢(m)←∑i=1nWi⁢(m)⁢xi⁢j⁢(Zi⁢(m)-Zˇi(j)⁢(m)), c1=ωj⁢(m)+t-ε-2⁢λ2⁢ε1+2⁢λ2 and c2=c12-4⁢(λ1-ωj⁢(m)⁢ε-t⁢ε1+2⁢λ2). Step 3.2: Update βj⁢(m) by Eq. (11); Step 4: Let m←m+1, β⁢(m+1)←β⁢(m); If β⁢(m) dose not convergence, then repeat Steps 2, 3.

4.Results

4.1Simulation

Here, we have conducted simulation research to measure the gene selection and classify the ability of the proposed method. Several penalized logistic model technologies are compared in the experiment: the Lasso method, the L1/2 method, the SCAD method, the elastic net method, and the L1-Net method. We follow Li’s work [17] to perform a simulation experiment; that is, a graph with 200 different transcription factors (TFs) is simulated. Ten 10 genes are regulated by a TF, which means the simulated network consists of 2,200. The dependent variable y is designated as a binary value of zero or 1, and is related to the first 4 TFs and their target genes.

Two models are suggested in the simulation. In every model, there were 200 instances: 100 training and 100 for testing.

In the first model, we assumed TF and its target played an activator or repressor role in the outcome variable:

β=(3,310,…,310⏟10,-3,-310,…,-310⏟10,5,510,…,510⏟10,-5,-510,…,-510⏟10,0,…,0)

In the second model, a TF could be both an activator and repressor at the same time, and the rest setting is similar to the first model:

β=(3,-310,-310,-310,310,…,310⏟7,-3,310,310,310,-310,…,-310⏟7,
 5,-510,-510,-510,510,…,510⏟7,-5,510,510,510,-510,…,-510⏟7,0,…,0)

The cross-validation (CV) technique has been widely used in parameter tuning. Here, we use a 10-CV method [21, 22] to identify the optimal tuning parameters for the training set. Genes with zero coefficients in the predicated model will be considered irrelevant to the predictor variables [23].

To consider the impact of variable correlation on the method more fully, a variable ρ was used to control the correlation between TF and its target.

The simulation process was repeated 500 times, and we use P and TP to report the feature selection ability of this method. P refers to the number of non-zero coefficient genes in the prediction model, and TP refers to the number of true non-zero coefficient genes in the model. The classification accuracy for the test set was also calculated, and Tables 1 and 2 summarize the results of each model.

Table 1

Simulation study – gene selection performance

LassoL1/2SCADElasticNetL1-NetLogsum-NL
ρ PTPPTPPTPPTPPTPPTP
Model 10.256.36.954.99.3387.625348.718.6134.724.1104.332.2
0.576.712.458.717.7449.733.7377.431.2130.224.8110.934.5
0.769.314.371.922.3499.141.5404.935.5165.630.8128.140.9
Model 20.253.77.9477.9366.415.2312.217.8109.432.894.134.7
0.560.110.250.612293.727.2339.923.9136.127121.139.1
0.761.810.561.310.9380.732.1372.230.4215.236.6128.238.3

Table 2

Simulation study – classification performance

LassoL1/2SCADElasticNetL1-NetLogsum-NL
ρ Accuracy
Model 10.285.47%82.43%82.10%84.64%85.85%92.93%
0.582.43%80.08%79.00%77.38%80.71%92.16%
0.777.96%81.33%78.93%79.79%81.44%91.54%
Model 20.287.67%89.60%88.41%86.00%89.21%95.18%
0.584.30%89.19%88.33%87.55%87.65%94.36%
0.781.34%85.52%84.82%84.09%86.50%92.63%

As shown in Table 1, compared to other algorithms, our method is more accurate at identifying real genes. For example, in Model 1, when ρ= 0.7, the mean TP identified by the Logsum-NL method is 40.9, while the number of true non-zero genes is 44. Our method selects almost entirely true genes. In addition, our method also performs well in the classification task. As dedicated in Table 2, our method has higher accuracy than sparse logistic regression models such as Lasso, L1/2, SCAD, elastic net, and L1-Net.

These results show that this method is a useful tool for classification and feature selection.

4.2Real data

To further prove the performance of the proposed method, we compared our approach with the other five regularization methods in an analysis of TCGA breast cancer. This data describes 20,501 genes in 806 different breast cancer samples. We retained only samples with complete information. After that, 85 TNBC and 460 non-TNBC were further divided into two groups: training (n= 327; 51 TNBC, 276 non-TNBC) and testing (n= 218; 34 TNBC, 184 non- TNBC) sets.

An extensive biological interactive network was obtained from BioGrid, which consists of 15,211 nodes (gene or other entities) and 336,119 interactions. A prepared network L with 11,320 genes and 224,458 edges was gained when we were mapping the downloaded network into the gene expression data.

We also added two new methods for performance comparison: SPL-Logsum [11] and HLR [7]. Table 3 shows that the Logsum-NL method gained higher predicting AUC performance than other mainstream regularization methods.

Table 3

The results for breast cancer

Method# selected genesTraining AUC (10-CV)Testing AUC
Lasso [1]6681.56%86.83%
L1/2 [6]5080.82%85.28%
SCAD [2]5487.67%79.94%
ElasticNet [16]8280.04%82.61%
L1-Net [24]13386.97%88.45%
HLR [7]8687.13%86.90%
SPL-Logsum [11]7987.26%88.71%
Logsum-NL9189.66%90.02%

Table 4

The top ten ranked genes

RankLassoL1/2SCADElasticNetL1-NetHLRSPL-LogsumLogsum-NL
1MYOM2CALCATMEM63BAKAP14ATP1A2CA12ATP1A2ESR1
2HOXB9FETUBWFDC1APCSGGA1FHOD1ABCA8SPSB1
3DCBLD1GABRB1DAPK3C15orf41KCNA2ETV4MYOM2DAPK3
4PGBD5PNMAL2SLC25A1GABRB1GABRDDAPK3FNBP1RTEL1
5MYOM2TM4SF4OVOL1KCNA2GBE1LHFPGATA3TM4SF4
6GABRB1PPP2R2DSLC27A4ATP1A2UPK3AGATA1TM4SF4EPHA8
7APCSADAMTS6MIFGABRDHEYLDAPK3CANT1GBE1
8USP9YKCNA2OPN4DCBLD1CPNE6UPK3AFOXO1PCDH10
9HSP90AA1ATP1A2GADD45BKIF15GUCY1A2ADAMTS6GBE1ATP1A2
10KCNA2FHOD1OLIG1HOXB9TAF9BSDC1ABI3BPLY6H

It can be seen from Table 4 that the genes identified by our Logsum-NL method include the SplA/Ryanodine Receptor Domain and SOCS Box intron 1 (SPSB 1), which has recently been identified as spontaneously regulated during breast tumor recurrence, and necessary and sufficient for promoting tumor recurrence [25]. The estrogen receptor (ESR 1) is one of the important markers for the classification of breast cancer subtypes in clinics, which can be used to not only guide prognosis but also decide treatment [26]. In breast tumors, protoprotein 10 (PCDH10) is down-regulated and methylated excessively [27]. The lymphocyte antigen 6 family member H (LY6H) is a cancer biomarker and therapeutic target that induces invasion and metastasis. LY6H is involved in the development of breast cancer by affecting the cellular pathway Ras/ERK. This gene may be a new marker for diagnosis and gene therapy in breast cancer patients [28].

5.Discussion and conclusions

The logsum method is a powerful method for feature selection. However, it is unable to use any previous biological structure information. To overcome this drawback, in this paper we first propose Logsum-Net regularization to integrate biological network knowledge. Then, we suggest the penalized logsum-net regularization logistic regression model (logsum-NL) for gene selection and cancer classification. For a real large dataset, the proposed method has achieved 89.66% (training) and 90.02% (testing) AUC performance which are, on average, 5.17% (training) and 4.49% (testing) better than mainstream methods. Therefore, the proposed logsum-NL method is a promising tool for gene selection and cancer classification of high-dimensional biological data. The limitation of this article is that it does not include an in-depth analysis of the selected genes. Future directions for research include further analysis of the potential clinical application of the selected genes, and investigation of the method with other high-dimensional data/models.

Acknowledgments

This research was supported by Macau Science and Technology Development Funds (0158/2019/A3, 0002/2019/APD) of Macau SAR of China, Special Innovation Projects of Universities in Guangdong Province (2018KTSCX205), the Natural Science Foundation of Guangdong Province (2018A030307033) and the National Natural Science Foundation of China (6201101081, 62006155).

Conflict of interest

None to report.

References

[1] 

Tibshirani R. Regression shrinkage and selection via the lasso. J R Stat Soc B. (1996) ; 267–88.

[2] 

Fan J, Li R. Variable selection via nonconcave penalized likelihood and its oracle properties. J Am Stat Assoc. (2001) ; 96: : 1348–60. doi: 10.1198/016214501753382273.

[3] 

Zou H. The adaptive lasso and its oracle properties. J Am Stat Assoc. (2006) ; 101: : 1418–29.

[4] 

Xu Z, Zhang H, Wang Y, Chang X, Liang Y. L1/2 regularization. Sci China Inf Sci. (2010) ; 53: : 1159–69.

[5] 

Huang H-H, Liang Y. Hybrid L1/2+2 method for gene selection in the Cox proportional hazards model. Comput Methods Programs Biomed. (2018) ; 164: : 65–73. doi: 10.1016/j.cmpb.2018.06.004.

[6] 

Liang Y, Liu C, Luan XZ, Leung KS, Chan TM, Xu ZB, et al. Sparse logistic regression with a L1/2 penalty for gene selection in cancer classification. BMC Bioinformatics. (2013) ; 14: : 198.

[7] 

Huang H-H, Liu X-Y, Liang Y. Feature selection and cancer classification via sparse logistic regression with the hybrid L1/2+2 regularization. PLoS One. (2016) ; 11: : e0149675. doi: 10.1371/journal.pone.0149675.

[8] 

Chu G-J, Liang Y, Wang J-X. Novel harmonic regularization approach for variable selection in Cox’s proportional hazards model. Comput Math Methods Med. (2014) ; 2014: : 857398. doi: 10.1155/2014/857398.

[9] 

Candès EJ, Wakin MB, Boyd SP. Enhancing sparsity by reweighted l 1 minimization. J Fourier Anal Appl. (2008) ; 14: : 877–905. doi: 10.1007/s00041-008-9045-x.

[10] 

Qiao B, Liu J, Liu J, Yang Z, Chen X. An enhanced sparse regularization method for impact force identification. Mech Syst Signal Process. (2019) ; 126: : 341–67. doi: 10.1016/J.YMSSP.2019.02.039.

[11] 

Xia L-Y, Wang Q-Y, Cao Z, Liang Y. Descriptor selection improvements for quantitative structure-activity relationships. Int J Neural Syst. (2019) ; 29: : 1950016. doi: 10.1142/S0129065719500163.

[12] 

Huang H, Liang Y. A novel Cox proportional hazards model for high-dimensional genomic data in cancer prognosis. IEEE/ACM Trans Comput Biol Bioinforma. (2019) : 1–1. doi: 10.1109/TCBB.2019.2961667.

[13] 

Huang HH, Liu XY, Li HM, Liang Y. Molecular pathway identification using a new L1/2 solver and biological network-constrained mode. Int J Data Min Bioinform. (2017) ; 17: : 189. doi: 10.1504/IJDMB.2017.085277.

[14] 

Huang H-H, Liang Y. An integrative analysis system of gene expression using self-paced learning and SCAD-Net. Expert Syst Appl. (2019) ; 135: : 102–12. doi: 10.1016/J.ESWA.2019.06.016.

[15] 

Huang H-H, Liang Y, Liu X-Y. Network-based logistic classification with an enhanced L1/2 solver reveals biomarker and subnetwork signatures for diagnosing lung cancer. Biomed Res Int. (2015) ; 2015: : 713953. doi: 10.1155/2015/713953.

[16] 

Zou H, Hastie T. Regularization and variable selection via the elastic net. J R Stat Soc Ser B. (2005) ; 67: : 301–20.

[17] 

Li C, Li H. Network-constrained regularization and variable selection for analysis of genomic data. Bioinformatics. (2008) ; 24: : 1175–82. doi: 10.1093/bioinformatics/btn081.

[18] 

Chen J, Zhang S, C. S. Integrative analysis for identifying joint modular patterns of gene-expression and drug-response data. Bioinformatics. (2016) ; 32: : 1724–32. doi: 10.1093/bioinformatics/btw059.

[19] 

Wang R, Su C, Wang X, Fu Q, Gao X, Zhang C, et al. Global gene expression analysis combined with a genomics approach for the identification of signal transduction networks involved in postnatal mouse myocardial proliferation and development. Int J Mol Med. (2018) ; 41: : 311–21. doi: 10.3892/ijmm.2017.3234.

[20] 

Friedman J, Hastie T, Tibshirani R. Regularization paths for generalized linear models via coordinate descent. J Stat Softw. (2010) ; 33: : 1–22.

[21] 

Chui K, Lytras M. A novel MOGA-SVM multinomial classification for organ inflammation detection. Appl Sci. (2019) ; 9: : 2284. doi: 10.3390/app9112284.

[22] 

Imai S, Takekuma Y, Kashiwagi H, Miyai T, Kobayashi M, Iseki K, et al. Validation of the usefulness of artificial neural networks for risk prediction of adverse drug reactions used for individual patients in clinical practice. PLoS One. (2020) ; 15: : e0236789. doi: 10.1371/journal.pone.0236789.

[23] 

Peng X, Yang Y. Algorithms for interval-valued fuzzy soft sets in stochastic multi-criteria decision making based on regret theory and prospect theory with combined weight. Appl Soft Comput. (2017) ; 54: : 415–30. doi: 10.1016/J.ASOC.2016.06.036.

[24] 

Ren J, Du Y, Li S, Ma S, Jiang Y, Wu C. Robust network-based regularization and variable selection for high-dimensional genomic data in cancer prognosis. Genet Epidemiol. (2019) ; 43: : 276–91. doi: 10.1002/gepi.22194.

[25] 

Feng Y, Pan T-C, Pant DK, Chakrabarti KR, Alvarez JV, Ruth JR, et al. SPSB1 promotes breast cancer recurrence by potentiating c-MET signaling. Cancer Discov. (2014) ; 4: : 790–803. doi: 10.1158/2159-8290.CD-13-0548.

[26] 

Koboldt DC, Fulton RS, McLellan MD, Schmidt H, Kalicki-Veizer J, McMichael JF, et al. Comprehensive molecular portraits of human breast tumours. Nature. (2012) ; 490: : 61–70. doi: 10.1038/nature11412.

[27] 

Li Z, Chim JCS, Yang M, Ye J, Wong BCY, Qiao L. Role of PCDH10 and its hypermethylation in human gastric cancer. Biochim Biophys Acta – Mol Cell Res. (2012) ; 1823: : 298–305. doi: 10.1016/J.BBAMCR.2011.11.011.

[28] 

Kong HK, Yoon S, Park JH. The regulatory mechanism of the LY6K gene expression in human breast cancer cells. J Biol Chem. (2012) ; 287: : 38889–900. doi: 10.1074/jbc.M112.394270.

Show more