1. 处理数据

samtools faidx genome.fasta #建索引
awk '$3=="dispersed_repeat" {print $1"\t"$4-1"\t"$5}' female_repeat.gff3 > TE.bed#TE.gff 转 BED，将注释文件中含有dispersed_repeat的行提取出来
bedtools makewindows -g genome.fasta.fai -w 100000 > windows.bed # 生成窗口（100kb，可改 50000 50kb）
bedtools coverage -a windows.bed -b TE.bed > TE_coverage.bed #统计密度
awk '{print $1"\t"$2"\t"$7}' TE_coverage.bed > TE_density.txt
 #TE_density.txt 三列：chr, start, density 备注：这一步当中包含有为挂载到染色体上的部分，画图前需要去除，如果没有去除就会出现以下报错：`geom_line()`: Each group consists of only one observation.i Do you need to adjust the group aesthetic?
grep "^Chr" TE_density.txt > TE_density_chr.txt #只保留染色体的行

2. 画图

library(circlize)

# 读取数据
dat <- read.table("TE_density.txt", sep = "\t", header = FALSE)

# 打开 PDF
pdf("TE_circos_clean.pdf", width = 9, height = 9)

circos.clear()
circos.par(
  start.degree = 90,
  gap.degree = 2,
  points.overflow.warning = FALSE
)

# 初始化染色体
circos.initialize(
  factors = dat$V1,
  xlim = cbind(0, tapply(dat$V2, dat$V1, max))
)

# 1. 染色体框（浅灰底色）
circos.track(
  ylim = c(0, 1),
  track.height = 0.08,
  bg.col = "#f2f2f2",
  bg.border = "#666666",
  panel.fun = function(x, y) {
    chr <- get.cell.meta.data("sector.index")
    x_mid <- get.cell.meta.data("xcenter")
    # 标注染色体名字
    circos.text(
      x_mid, 0.5, chr,
      facing = "bending", nice.facing = TRUE,
      cex = 0.9, col = "black", font = 2
    )
  }
)

# 2. TE 密度曲线（红色）
circos.track(
  ylim = c(0, 1),
  track.height = 0.4,
  panel.fun = function(x, y) {
    chr <- get.cell.meta.data("sector.index")
    sub <- dat[dat$V1 == chr, ]
    # 密度折线
    circos.lines(sub$V2, sub$V3, col = "#e63946", lwd = 1.5)
    # 中间参考线
    circos.lines(c(0, max(sub$V2)), c(0.5, 0.5), col = "gray", lty = 3, lwd = 0.8)
  }
)

dev.off()
circos.clear()