[英]R: Using Log Rank Test (survdiff)
OK, so I have a dataframe that looks like this: 好的,所以我有一个像这样的数据框:
head(exprs, 21)
sample expr ID X_OS
1 BIX high TCGA_DM_A28E_01 26
2 BIX high TCGA_AY_6197_01 88
3 BIX high TCGA_HB_KH8H_01 553
4 BIX low TCGA_K4_6303_01 256
5 BIX low TCGA_F4_6703_01 491
6 BIX low TCGA_Y7_PIK2_01 177
7 BIX low TCGA_A6_5657_01 732
8 HEF high TCGA_DM_A28E_01 26
9 HEF high TCGA_AY_6197_01 88
10 HEF high TCGA_F4_6703_01 491
11 HEF high TCGA_HB_KH8H_01 553
12 HEF low TCGA_K4_6303_01 256
13 HEF low TCGA_Y7_PIK2_01 177
14 HEF low TCGA_A6_5657_01 732
15 TUR high TCGA_DM_A28E_01 26
16 TUR high TCGA_F4_6703_01 491
17 TUR high TCGA_Y7_PIK2_01 177
18 TUR low TCGA_K4_6303_01 256
19 TUR low TCGA_AY_6197_01 88
20 TUR low TCGA_HB_KH8H_01 553
21 TUR low TCGA_A6_5657_01 732
Simply, for each sample
, there are 7 patients, each with a survival time ( X_OS
) and expression level high
or low
( expr
). 简单来说,对于每个
sample
,有7位患者,每位患者的生存时间( X_OS
)和表达水平high
或low
( expr
)。 In the code below, I wish to take the first sample and run it through the survdiff
function, with the outputs going to dfx
. 在下面的代码中,我希望获取第一个示例并通过
survdiff
函数运行它,输出将进入dfx
。 However, I'm new to survival analysis and I'm not sure how to use the parameters of the survdiff
function. 但是,我是生存分析的新手,我不确定如何使用
survdiff
函数的参数。 I wish to compare high
and low
expression groups for each sample
. 我希望比较每个
sample
high
表达组和low
表达组。 How can I edit the function expfun
to yield the survdiff
output I need? 如何编辑函数
expfun
以产生所需的survdiff
输出? In addition, ideally I'd love to get the pvalues out of it, but I can work on that in a later step. 另外,理想情况下,我很乐意从中获取pvalue,但是我可以在以后的步骤中进行研究。 Thank you!
谢谢!
expfun = function(x) {
survdiff(Surv(x$X_OS, x$expr))
}
dfx <- pblapply(split(exprs[c("expr", "X_OS")], exprs$sample), expfun)
Try this. 尝试这个。 I added a proper Surv() call because you only had times and no status argument and I made it into a formula (with the predictor on the other side of the tilde) because
Surv
function expects status as its second argument and survdiff
expects a formula as its first argument. 我添加了一个适当的Surv()调用,因为您只有时间,没有状态参数,并且将其放入公式中(预测变量位于波浪号的另一侧),因为
Surv
函数将状态期望为第二个参数,而survdiff
期望使用公式作为第一个论点。 That means you need to use the regular R regression calling convention where column names are used as the formula tokens and the dataframe is given to the data argument. 这意味着您需要使用常规的R回归调用约定,其中将列名用作公式标记,并将数据框指定给data参数。 If you had a censoring variable, it would be put in as the second Surv argument rather than the
1
's that I have in there now. 如果您有一个检查变量,它将作为第二个Surv参数而不是我现在在其中的
1
输入。
expfun = function(x) {
survdiff( Surv( X_OS, rep(1,nrow(x)) ) ~ expr, data=x)
}
dfx <- lapply(split(exprs[c("expr", "X_OS")], exprs$sample), expfun)
This is the result from print.survdiff: 这是来自print.survdiff的结果:
> dfx
$BIX
Call:
survdiff(formula = Surv(X_OS, rep(1, nrow(x))) ~ expr, data = x)
N Observed Expected (O-E)^2/E (O-E)^2/V
expr=high 3 3 2.05 0.446 0.708
expr=low 4 4 4.95 0.184 0.708
Chisq= 0.7 on 1 degrees of freedom, p= 0.4
$HEF
Call:
survdiff(formula = Surv(X_OS, rep(1, nrow(x))) ~ expr, data = x)
N Observed Expected (O-E)^2/E (O-E)^2/V
expr=high 4 4 3.14 0.237 0.51
expr=low 3 3 3.86 0.192 0.51
Chisq= 0.5 on 1 degrees of freedom, p= 0.475
$TUR
Call:
survdiff(formula = Surv(X_OS, rep(1, nrow(x))) ~ expr, data = x)
N Observed Expected (O-E)^2/E (O-E)^2/V
expr=high 3 3 1.75 0.902 1.41
expr=low 4 4 5.25 0.300 1.41
Chisq= 1.4 on 1 degrees of freedom, p= 0.235
Note that you can see the code to produce the print output with: 请注意,您可以使用以下代码查看生成打印输出的代码:
getAnywhere(print.survdiff)
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.