我们正试图在四个不同的城市:BKI、Daus、SRINA和Palla寻找土壤有毒元素与表观遗传年龄加速之间的联系.利用电感耦合等ionic 体质谱(ICPMS)对土壤中的有毒元素进行了测定,并与年代进行了回归,计算了表观遗传年龄加速度.我们总共使用了10种有毒元素.
然而,我面临着两个问题: 1.所有四个城市的参与者人数各不相同:BKI(1000名参与者)、Daus(250名参与者)、SRINA(200名参与者)、Palla(100名参与者).我应该如何解释这些不同数量的个人?
2.我们想在每个城市测量所有10种有毒元素与表观遗传年龄的关联.
目前,我正在使用下面的模型,但它提供了所有元素和表观遗传年龄加速之间的关联.我们还希望根据城市将yields 分开.
library("QuantPsyc")
data <- read.table("clr.clean.file2.txt", header=T, sep=",")
model1 <- lm (epigenetic_age_acceleration ~
As + Se + Fe + Co + Zn + Mn + Hg + Sb + Mo + Pb +
Smoking_Status + Sex + Age, data = data)
model1
model1.stat<-lm.beta (model1)
model1.stat
As
0.0256056478741109
Se
0.00499178037586947
Fe
0.00210283404005497
Co
-0.00916637143431217
Zn
0.0639371964557919
Mn
-0.0213600659139311
Hg
0.0328431516176923
Sb
0.000169338014091565
Mo
-0.0200956999960768
suggested code work of me;
IEAA_elements <- lm (epigentic_age_acceleration ~ 0 + city +
city:(As +Se + Fe + Co + Zn +Mn + Hg + Sb+ Mo +Pb +
Smoking_Status + Sex + Age), data = data)
cf <- confint(IEAA_elements)
cf
现在不确定如何将置信度区间与回归输出合并?
Also not sure if I have to use GLM model then should I use gaussian family but not sure about link type? and what role does it has in GLM model
# Fit a GLM
formula <- as.formula(epigentic_age_acceleration ~ 0 + city +
city:(As +Se + Fe + Co + Zn +Mn + Hg + Sb+ Mo +Pb +
Smoking_Status + Sex + Age))
model <- glm(formula, data = data, family = gaussian(link="identity"))