非癌症也能做逐步，logistic逐步回归
尔云间 一个专门做科研的团队
小果最近在做非疾病的分析，其中就涉及到了逐步回顾的分析，之前小果做过癌症的逐步回归，但没做过非癌症的，于是小果去找了一下逐步的代码，下面就让我们一起来看看吧。
代码如下：
setwd("C:/Users/Administrator/Desktop")
data=read.table("逐步.txt",header = T,row.names = 1,sep = "\t")
library(gtsummary)
fitMul<-glm(group~.,data=data,family=binomial())
fitMul<- step(fitMul,direction = "both") #下面为逐步分析结果
Start: AIC=92.88
group ~ AURKA + EHD1 + HTRA2 + LIAS + MRPS14 + MYC + NFU1 + NNMT +
SMYD2 + TACO1 + UBE2T
Df Deviance AIC
- LIAS 1 68.881 90.881
- TACO1 1 68.897 90.897
- MRPS14 1 68.975 90.975
- SMYD2 1 69.005 91.005
- EHD1 1 70.130 92.130
<none> 68.879 92.879
- NFU1 1 72.151 94.151
- HTRA2 1 72.170 94.170
- NNMT 1 72.207 94.207
- UBE2T 1 79.043 101.043
- AURKA 1 84.408 106.408
- MYC 1 99.722 121.722
Step: AIC=90.88
group ~ AURKA + EHD1 + HTRA2 + MRPS14 + MYC + NFU1 + NNMT + SMYD2 +
TACO1 + UBE2T
Df Deviance AIC
- TACO1 1 68.910 88.910
- MRPS14 1 68.981 88.981
- SMYD2 1 69.010 89.010
- EHD1 1 70.138 90.138
<none> 68.881 90.881
- NFU1 1 72.167 92.167
- NNMT 1 72.231 92.231
- HTRA2 1 72.371 92.371
+ LIAS 1 68.879 92.879
- UBE2T 1 79.443 99.443
- AURKA 1 84.435 104.435
- MYC 1 99.735 119.735
Step: AIC=88.91
group ~ AURKA + EHD1 + HTRA2 + MRPS14 + MYC + NFU1 + NNMT + SMYD2 +
UBE2T
Df Deviance AIC
- SMYD2 1 69.085 87.085
- MRPS14 1 69.119 87.119
- EHD1 1 70.144 88.144
<none> 68.910 88.910
- NFU1 1 72.231 90.231
- NNMT 1 72.566 90.566
- HTRA2 1 72.595 90.595
+ TACO1 1 68.881 90.881
+ LIAS 1 68.897 90.897
- UBE2T 1 79.614 97.614
- AURKA 1 85.261 103.261
- MYC 1 99.994 117.994
Step: AIC=87.08
group ~ AURKA + EHD1 + HTRA2 + MRPS14 + MYC + NFU1 + NNMT + UBE2T
Df Deviance AIC
- MRPS14 1 69.276 85.276
- EHD1 1 70.180 86.180
<none> 69.085 87.085
- NFU1 1 72.705 88.705
- NNMT 1 72.782 88.782
+ SMYD2 1 68.910 88.910
+ TACO1 1 69.010 89.010
+ LIAS 1 69.054 89.054
- HTRA2 1 74.058 90.058
- UBE2T 1 79.624 95.624
- AURKA 1 88.230 104.230
- MYC 1 101.523 117.523
Step: AIC=85.28
group ~ AURKA + EHD1 + HTRA2 + MYC + NFU1 + NNMT + UBE2T
Df Deviance AIC
- EHD1 1 70.369 84.369
<none> 69.276 85.276
- NNMT 1 72.970 86.970
+ TACO1 1 69.076 87.076
+ MRPS14 1 69.085 87.085
+ SMYD2 1 69.119 87.119
+ LIAS 1 69.183 87.183
- NFU1 1 74.035 88.035
- HTRA2 1 77.191 91.191
- UBE2T 1 81.170 95.170
- AURKA 1 88.866 102.866
- MYC 1 103.781 117.781
Step: AIC=84.37
group ~ AURKA + HTRA2 + MYC + NFU1 + NNMT + UBE2T
Df Deviance AIC
<none> 70.369 84.369
+ EHD1 1 69.276 85.276
+ MRPS14 1 70.180 86.180
- NFU1 1 74.240 86.240
+ TACO1 1 70.339 86.339
+ SMYD2 1 70.342 86.342
+ LIAS 1 70.365 86.365
- NNMT 1 76.817 88.817
- HTRA2 1 78.932 90.932
- UBE2T 1 82.157 94.157
- AURKA 1 89.783 101.783
- MYC 1 106.493 118.493
#下面继续分析代码
fitSum<-summary(fitMul)
ResultMul<-c()#准备空向量，用来储存结果
ResultMul<-rbind(ResultMul,fitSum$coef)
OR<-exp(fitSum$coef[,'Estimate'])
ResultMul<-cbind(ResultMul,cbind(OR,exp(confint(fitMul))))
#Waiting for profiling to be done...
ResultMul #查看分析结果
Estimate Std. Error z value Pr(>|z|) OR 2.5 % 97.5 %
(Intercept) -51.047887 12.5263722 -4.075233 4.596828e-05 6.763703e-23 6.811140e-35 3.443089e-13
AURKA 5.086038 1.3770334 3.693475 2.212104e-04 1.617478e+02 1.354446e+01 3.198491e+03
HTRA2 3.541670 1.3155403 2.692179 7.098684e-03 3.452452e+01 3.069880e+00 5.844779e+02
MYC -2.410618 0.5051995 -4.771617 1.827532e-06 8.975980e-02 2.951109e-02 2.198625e-01
NFU1 2.560629 1.3508181 1.895614 5.801116e-02 1.294396e+01 1.009701e+00 2.159337e+02
NNMT -1.345624 0.5717925 -2.353344 1.860544e-02 2.603771e-01 7.681164e-02 7.461131e-01
UBE2T 2.581739 0.8298538 3.111077 1.864062e-03 1.322011e+01 2.880822e+00 7.776186e+01
这个分析和单因素的分析差不多，小伙伴们如果知道单因素的分析方法的话，那这个肯定一看就懂。
好了，这就是今天的主要内容了，小伙伴们有什么问题欢迎来和小果分享讨论啊。
shengxinguoer
生信果
生信硬核知识解答
和小果一起学生信
