>我有一个数据帧sampleManifest
,如下所示:
SampleName Status Role Sex
AU056001_00HI1299A unaffected sibling female
AU056002_00HI1301A unaffected proband male
AU0780201_00HI1775A unaffected father male
AU0780202_00HI1777A unaffected mother female
AU0780301_00HI1778A affected proband male
.
.
.
以及成对样本比较的单独数据框,kinshipEstimates
:
FID ID1 ID2 Kinship Relationship
AU0560 AU056001_00HI1299A AU056002_00HI1301A 0.0283 full-sibling
AU0780 AU0780201_00HI1775A AU0780202_00HI1777A -0.00160 unrelated
AU0780 AU0780201_00HI1775A AU0780301_00HI1778A 0.284 parent-child
AU0780 AU0780202_00HI1777A AU0780301_00HI1778A 0.246 parent-child
.
.
.
我想构建一个新的数据帧,其中每行kinshipEstimates
中两个样本中的每一个的sampleManifest$Role
,使其看起来像这样:
FID ID1 ID2 Roles Kinship Relationship
AU0560 AU056001_00HI1299A AU056002_00HI1301A sibling-proband 0.0283 full-sibling
AU0780 AU0780201_00HI1775A AU0780202_00HI1777A father-mother -0.00160 unrelated
AU0780 AU0780201_00HI1775A AU0780301_00HI1778A father-proband 0.284 parent-child
AU0780 AU0780202_00HI1777A AU0780301_00HI1778A mother-proband 0.246 parent-child
.
.
.
我一直在尝试使用left_join
,但不知道如何将成对中每个样本的相应Role
合并为一个值。
解决方案是使用双left_join
tidyverse
包。首先加入kinshipEstimates
sampleManifest
ID1
和SampleName
。再次加入sampleManifest
,结果在ID2
和SampleName
上。最后,使用tidyr::unite
合并Role.x
和Role.y
。
library(tidyverse)
left_join(kinshipEstimates, sampleManifest, by=c("ID1" = "SampleName")) %>%
select(-Status, -Sex) %>%
left_join(sampleManifest, by=c("ID2" = "SampleName")) %>%
unite(Roles, Role.x, Role.y, sep="-") %>%
select(-Sex, -Status)
# FID ID1 ID2 Kinship Relationship Roles
# 1 AU0560 AU056001_00HI1299A AU056002_00HI1301A 0.0283 full-sibling sibling-proband
# 2 AU0780 AU0780201_00HI1775A AU0780202_00HI1777A -0.0016 unrelated father-mother
# 3 AU0780 AU0780201_00HI1775A AU0780301_00HI1778A 0.2840 parent-child father-proband
# 4 AU0780 AU0780202_00HI1777A AU0780301_00HI1778A 0.2460 parent-child mother-proband
数据:
sampleManifest <- read.table(text =
"SampleName Status Role Sex
AU056001_00HI1299A unaffected sibling female
AU056002_00HI1301A unaffected proband male
AU0780201_00HI1775A unaffected father male
AU0780202_00HI1777A unaffected mother female
AU0780301_00HI1778A affected proband male",
stringsAsFactors = FALSE, header = TRUE)
kinshipEstimates <- read.table(text =
"FID ID1 ID2 Kinship Relationship
AU0560 AU056001_00HI1299A AU056002_00HI1301A 0.0283 full-sibling
AU0780 AU0780201_00HI1775A AU0780202_00HI1777A -0.00160 unrelated
AU0780 AU0780201_00HI1775A AU0780301_00HI1778A 0.284 parent-child
AU0780 AU0780202_00HI1777A AU0780301_00HI1778A 0.246 parent-child",
stringsAsFactors = FALSE, header = TRUE)
这是一种使用gather
、 一个inner_join
和group_by
的方法。
添加行号可以让我们在分组时跟踪ID1
/ID2
对:
kinshipEstimates %>%
mutate(row_num = row_number()) %>%
gather(which_id, id, -row_num, -FID, -Kinship, -Relationship) %>%
inner_join(sampleManifest, by=c("id" = "SampleName")) %>%
group_by(FID, row_num) %>%
summarise(Roles = paste(Role, collapse="-"),
Kinship = first(Kinship),
Relationship = first(Relationship))
FID row_num Roles Kinship Relationship
<chr> <int> <chr> <dbl> <chr>
1 AU0560 1 sibling-proband 0.0283 full-sibling
2 AU0780 2 father-mother -0.00160 unrelated
3 AU0780 3 father-proband 0.284 parent-child
4 AU0780 4 mother-proband 0.246 parent-child