ForestPlot

博主： int0030
发布时间：2024 年 03 月 01 日
394 次浏览
2 条评论
1115字数
分类： wiki

# forestploter

## install

```R
library(forestploter)
```

## format

![table format](https://img.int0030.xyz/images/2024/03/01/77fd092f07b5687f12089b8b5c60.png)

参照以上表格，必须的列包括：

1. Exposure：包括暴露的名字、MR方法等
2. OR(95% CI)：使用or列、lo_ci、up_ci列生成
3. snps、P-Value
4. se.or、lo_ci、up_ci、or：用于绘图

## Example

```R
setwd("/Users/luojy/同步/学位论文/2.MR筛选出来的各暴露VSAS结果/")
library(dplyr)
tmp <- readxl::read_excel("1.MR显著.xlsx", col_types = c("text","text","text","text","numeric","text","text","text","text","text","text","text","text","numeric","numeric","numeric","numeric","numeric","numeric","numeric","text","text","numeric","numeric"))

lipid_res <- tmp %>% 
    dplyr::select(filename, exposure_factor_abbr, exposure_factor_class,exposure_factor_main_class, exposure_factor_sub_class,method,nsnp, b.or, se.or, pval.or, or, lo_ci, up_ci,N) %>% 
    dplyr::filter(exposure_factor_class == 'Lipid') %>% 
    unique() %>%
    group_by(exposure_factor_abbr) %>%
    dplyr::filter(!(method == "TwoSampleMR" & "Inverse variance weighted" %in% method[exposure_factor_abbr == .data$exposure_factor_abbr])) %>%
    dplyr::mutate(method = ifelse(method == "TwoSampleMR", "Inverse variance weighted", method)) %>% 
    ungroup()

## Lipid中，IVW下的脂质森林图，按照exposure_factor_main_class和exposure_factor_sub_class脂质类型分类
library(forestploter)
tibble_insert_row <- function(df, row, pos) {
    if (pos > nrow(df)) {
        return(rbind(df, row))
    }
    if (pos == 1) {
        return(rbind(row, df))
    }
    return(rbind(df[1:(pos-1),], row, df[pos:nrow(df),]))
}

split_forest_data_by_singletitle <- function(data, input_string) {
  # 查找匹配行的索引
  start_index <- which(data$Exposure == input_string)
  
  # 初始化结束索引为数据框的行数
  end_index <- nrow(data)
  
  # 查找下一个同级别分类的索引作为结束索引
  for (i in (start_index + 1):nrow(data)) {
    # 如果找到下一个主分类，则更新结束索引并停止循环
    if (data$Exposure[i] != "" & substr(data$Exposure[i], 1, 2) != "  ") {
      end_index <- i - 1
      break
    }
  }
  
  # 提取匹配行及其子项
  return(data[start_index:end_index, ])
}

split_forest_data_by_title <- function(data, input_string) {
    for (i in 1:length(input_string)){
        sub_data <- split_forest_data_by_singletitle(data, input_string[i])
        if (i == 1){
            result <- sub_data
        }else{
            result <- rbind(result, sub_data)
        }
    }
    return(result)
}

forestplot_by_title <- function(data, title,pdf_output = NULL) {
  # 提取匹配行及其子项
  sub_data <- split_forest_data_by_title(data, title)
  
  # 提取OR和CI
  ci <- sub_data %>% dplyr::select(or, lo_ci, up_ci, se.or)
  
  # 提取需要的列
  sub_data <- sub_data %>% 
    dplyr::select(Exposure, method, snps, `OR(95% CI)`, ` `, `P-Value`)
  
  # 绘制森林图
  p <- forest(sub_data,
         est = as.numeric(ci$or),       #效应值
         lower = as.numeric(ci$lo_ci),     #可信区间下限
         upper = as.numeric(ci$up_ci),      #可信区间上限
         sizes = as.numeric(ci$se.or),
         ci_column = 5,   #在那一列画森林图，要选空的那一列
         ref_line = 1,
         arrow_lab = c("Negtive", "Posive"),
         xlim = c(0.8, 1.2),
         ticks_at = c(0.8, 1, 1.2),)

if (is.null(pdf_output)){
        message("No pdf output")
        print(p)
    } else {
        if(nrow(sub_data)<10){
            height <- nrow(sub_data) * 0.5
        } else if (nrow(sub_data)<5){
            height <- 5
        }
        else {
            height <- nrow(sub_data) * 0.3
        }
  
        pdf(pdf_output, width = 12, height = height)
        print(p)
        dev.off()
    }
}

forest_data <- lipid_res %>%
    dplyr::arrange(exposure_factor_main_class, exposure_factor_sub_class) %>% 
    dplyr::select(exposure_factor_main_class,method, exposure_factor_sub_class,N, or,pval.or, se.or,nsnp,exposure_factor_abbr) %>% 
    dplyr::mutate(exposure_factor_abbr = paste0("  ",exposure_factor_abbr)) %>%
    dplyr::mutate(title = paste0(exposure_factor_main_class,": ",exposure_factor_sub_class)) %>%
    dplyr::mutate(lo_ci = or-se.or*1.96) %>%
    dplyr::mutate(up_ci = or+se.or*1.96) %>%
    dplyr::mutate(`95%CI` = paste(sprintf("%.2f (%.2f to %.2f)",or, lo_ci,up_ci))) %>%
    dplyr::mutate(pval.or = round(as.numeric(pval.or), 4)) %>% 
    dplyr::select(exposure_factor_main_class, exposure_factor_sub_class, method, N , `95%CI`, or, pval.or, se.or,lo_ci,up_ci,nsnp, exposure_factor_abbr, title)

tmp <- forest_data %>% 
    dplyr::select(title) %>% 
    unique() %>% 
    dplyr::mutate(first_occurrences = sapply(., function(x) match(x, forest_data$title)))

forest_data_tmp <- forest_data
for (i in nrow(tmp):1){
    forest_data_tmp <- forest_data_tmp %>% 
        tibble_insert_row(c(rep('',11),tmp[i,1][[1]][[1]],rep('',ncol(forest_data)-11-1)), tmp[i,2][[1]][[1]])
}

forest_data_tmp <- forest_data_tmp %>% dplyr::mutate(" "=paste(rep(" ",15),collapse = " "))

ci <- forest_data_tmp %>% dplyr::select(or,lo_ci, up_ci, se.or)

forest_data_tmp <- forest_data_tmp %>% 
    dplyr::rename(Exposure = exposure_factor_abbr, snps = nsnp, `OR(95% CI)` = `95%CI`, `P-Value` = pval.or)

title_list <- list(
    list("Fatty Acyls",c("Fatty Acyls: Dicarboxylic acids","Fatty Acyls: FA","Fatty Acyls: Fatty acyl carnitines","Fatty Acyls: Heterocyclic fatty acids","Fatty Acyls: Hydroxy fatty acids","Fatty Acyls: N-acyl amino acids","Fatty Acyls: UFA")),
    list("GL",c("GL: DG","GL: TG")),
    list("PL",c("PL: LPC","PL: LPE","PL: LPI","PL: PC","PL: PE","PL: PI","PL: PS","PL: aLPC","PL: aPC","PL: aPE","PL: pLPC","PL: pLPE","PL: pPC","PL: pPE")),
    list("PR",c("PR: Isoprenoids")),
    list("SL",c("SL: Cer","SL: GM3","SL: Hex2Cer","SL: Hex3Cer","SL: HexCer","SL: Phosphonosphingolipids","SL: SM")),
    list("Sterols",c("Sterols: Bile acids and derivatives","Sterols: CE","Sterols: Cholesterol","Sterols: Cholesterol and derivatives","Sterols: Ergosterols and C24-methyl derivatives","Sterols: Steroids","Sterols: Taurine conjugates"))
)

for (i in title_list){
    pdf_file_path <- paste0(i[[1]],".pdf")
    p <- forestplot_by_title(forest_data_tmp,i[[2]],pdf_output = pdf_file_path)
    print(p)
    dev.off()
}

```

如图：

![Forest Plot Example](https://img.int0030.xyz/images/2024/03/01/1197d3b48bcbcc5e7f95d8c76502.png)

最后修改：2024 年 03 月 01 日

如果觉得我的文章对你有用，请随意赞赏

2 条评论

rechlgpbql
February 28th, 2025 at 09:19 pm

文化底蕴深厚，引经据典信手拈来。

回复
tblkbxdlii
February 28th, 2025 at 04:31 pm

文字流畅如丝，语言优美动人，读来令人心旷神怡。

回复

发表评论取消回复
使用cookie技术保留您的个人信息以便您下次快速评论，继续评论表示您已同意该条款

评论 *

私密评论

名称 *

🎲

邮箱 *

地址

推流测试
浏览次数: 1310
第3章糖类
浏览次数: 1100
第九章色素和着色剂
浏览次数: 851
第二章水分
浏览次数: 706
实验室安全管理
浏览次数: 642

iczikzrcyo
这篇文章不错！
bzbjufqpdh
这篇文章不错！
xeizkhadzl
这篇文章不错！
wjijbdieri
这篇文章不错！
zrhvujjvxy
内容的丰富性和深度让人仿佛置身于知识的海洋，受益匪浅。

丑石
浏览次数: 202
丑石
浏览次数: 200
Lipid145-server | Gitlab | SSL | Docker部署 | GitlabRunner | 邮件
浏览次数: 203
UpSetR
浏览次数: 490
济南的冬天
浏览次数: 275

ForestPlot

int0030 • 2024 年 03 月 01 日

# forestploter

## install

```R
library(forestploter)
```

## format

![table format](https://img.int0030.xyz/images/2024/03/01/77fd092f07b5687f12089b8b5c60.png)

参照以上表格，必须的列包括：

1. Exposure：包括暴露的名字、MR方法等
2. OR(95% CI)：使用or列、lo_ci、up_ci列生成
3. snps、P-Value
4. se.or、lo_ci、up_ci、or：用于绘图

## Example

tmp <- forest_data %>% 
    dplyr::select(title) %>% 
    unique() %>% 
    dplyr::mutate(first_occurrences = sapply(., function(x) match(x, forest_data$title)))

forest_data_tmp <- forest_data_tmp %>% dplyr::mutate(" "=paste(rep(" ",15),collapse = " "))

ci <- forest_data_tmp %>% dplyr::select(or,lo_ci, up_ci, se.or)

forest_data_tmp <- forest_data_tmp %>% 
    dplyr::rename(Exposure = exposure_factor_abbr, snps = nsnp, `OR(95% CI)` = `95%CI`, `P-Value` = pval.or)

for (i in title_list){
    pdf_file_path <- paste0(i[[1]],".pdf")
    p <- forestplot_by_title(forest_data_tmp,i[[2]],pdf_output = pdf_file_path)
    print(p)
    dev.off()
}

```

如图：

![Forest Plot Example](https://img.int0030.xyz/images/2024/03/01/1197d3b48bcbcc5e7f95d8c76502.png)

2 条评论

发表评论 取消回复 使用cookie技术保留您的个人信息以便您下次快速评论，继续评论表示您已同意该条款

ForestPlot

发表评论取消回复
使用cookie技术保留您的个人信息以便您下次快速评论，继续评论表示您已同意该条款