在 GenomicRanges 对象中合并具有相同 属性 的相邻容器

Merging adjacent bins with same property in GenomicRanges object

将基因组分割成相邻的非重叠箱,例如通过 tileGenome,我通过某种方式为每个 bin(比如 1 或 2)计算了一些 属性。

现在我想合并相邻的相同 属性。 一个最小的例子如下所示:

library(GenomicRanges)
chrSizes <- c(chr1 = 1000, chr2 = 500)
bins   <- tileGenome(chrSizes, tilewidth = 200, cut.last.tile.in.chrom = T)
bins$property <- rep(1:2, each = 4)
bins
GRanges object with 8 ranges and 1 metadata column:
      seqnames    ranges strand |  property
         <Rle> <IRanges>  <Rle> | <integer>
  [1]     chr1     1-200      * |         1
  [2]     chr1   201-400      * |         1
  [3]     chr1   401-600      * |         1
  [4]     chr1   601-800      * |         1
  [5]     chr1  801-1000      * |         2
  [6]     chr2     1-200      * |         2
  [7]     chr2   201-400      * |         2
  [8]     chr2   401-500      * |         2
  -------
  seqinfo: 2 sequences from an unspecified genome

前 4 个 bin 有 属性 1,因此应合并为一个 bin。

我查看了 GRanges 文档,但找不到明显的本机解决方案。 请注意,必须考虑 seqname 边界(例如,chr1 和 chr2 保持分开,与 属性 无关) 显然,我可以使用循环,但我宁愿使用原生的 GRange 解决方案,例如使用我可能已经监督过的 union

所需的输出应如下所示:

      seqnames    ranges strand |  property
         <Rle> <IRanges>  <Rle> | <integer>
  [1]     chr1     1-800      * |         1
  [2]     chr1  801-1000      * |         2
  [3]     chr2     1-500      * |         2

R 基因组范围:

result <- unlist(reduce(split(bins, ~property)))
result$property <- names(result)

# GRanges object with 3 ranges and 1 metadata column:
# seqnames    ranges strand |    property
# <Rle> <IRanges>  <Rle> | <character>
# 1     chr1     1-800      * |           1
# 2     chr1  801-1000      * |           2
# 2     chr2     1-500      * |           2
# -------
# seqinfo: 2 sequences from an unspecified genome

Python PyRanges:

import pandas as pd
from io import StringIO
import pyranges as pr

c = """Chromosome Start End Value
chr1 1 200 Python
chr1 201 400 Python
chr1 401 600 Python
chr1 601 800 Python
chr1 801 1000 R
chr2 1 200 R
chr2 201 400 R
chr2 401 500 R"""

df = pd.read_table(StringIO(c), sep=" ")
gr = pr.PyRanges(df)
gr.merge(by="Value", slack=1)

# +--------------+-----------+-----------+------------+
# | Chromosome   |     Start |       End | Value      |
# | (category)   |   (int32) |   (int32) | (object)   |
# |--------------+-----------+-----------+------------|
# | chr1         |         1 |       800 | Python     |
# | chr1         |       801 |      1000 | R          |
# | chr2         |         1 |       500 | R          |
# +--------------+-----------+-----------+------------+
# Unstranded PyRanges object has 3 rows and 4 columns from 2 chromosomes.
# For printing, the PyRanges was sorted on Chromosome.