COXPH的变量筛选
返工的血泪教训，单因素COXPH做完显著是不够的，此时应该再来一次多因素COXPH，剔除那些不具有独立预后能力的变量。 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263library("survival&
COXPH可视化
建模123456789101112131415library(rms)library(survival)dat <- readRDS('../../../COXPH/data/PRAD.rds')pfg <- readRDS('../A_ref_A_fiig.3_A_B/pfg.rds')y <- Surv(round(dat$meta$p
LASSO回归：使用L1正则化控制过拟合
安装补充包 conda activate ggsurvplot conda install -c conda-forge r-glmnet -y LASSO回归12345678910library(glmnet)library(survival)dat <- readRDS('../../../COXPH/data/PRAD.rds')y <- Surv(roun
格里森评分数据处理
123456789101112f_TCGA_gleason_grade <- function(primary_gleason_grade, secondary_gleason_grade){ primary_gleason_grade <- as.numeric(unlist(data.frame(strsplit(primary_gleason_grade, &#x
ggcorrplot绘制相关性热图
安装包 conda activate wgcna conda install -c conda-forge r-ggcorrplot -y conda install -c conda-forge r-ggsci -y # conda install -c conda-forge r-gridextra -y 颜色参考12345678library("reshape2")se
使用NMF进行聚类
安装包 conda activate wgcna conda install -c conda-forge r-nmf -y 确定秩1234library(NMF)dat <- readRDS('PRAD.rds')res <- nmf(dat,2:7,nrun=10, seed=123)pdf(file = 'nmf_sp.pdf', wid
RRA结果绘图
安装包 # conda activate rplot # conda env remove -n ggVennDiagram # conda env remove -n rsf conda create -n rsf -c conda-forge r-sf=1.0_4 conda activate rsf library(sf) # install.packages(“sf”, vers
不同批次但未记录批次时使用DESeq2的方法
准备数据1234567891011121314151617181920212223242526library(SummarizedExperiment)tmp <- load('PRAD_tp.rda')counts <- data@assays@data$unstrandedcolnames(counts) <- data@colData$patientf_