将数据框与另一个数据框中的权重相乘

Question

我有一个带有表达式值的数据框df，我在数据框Weights中具有权重。 对于df中的每一列，我想将df的每一行与具有相似行名的相应行相乘Weights。

然后，对于df中的每一列，您将获得行的加权值。

请看我的输出。

东风

Gene              MMRF_1021    MMRF_1024   MMRF_1029   MMRF_1030    MMRF_1031
ENSG00000007062   0.05374547   0.01258559   0.0000000   1.2985088   0.37618693
ENSG00000012124   0.13436368   0.27688288   0.2780448   0.7158432   0.03271195

权重

Gene                   Pre.BI       Pre.BII       Immature     Naive         Memory       Plasmacell
ENSG00000007062        0.006368928  0.000000e+00  0.000000000  0.0000000000  0.000000000  0.000000000
ENSG00000012124        0.000000000  0.000000e+00  0.000000000  0.0000000000  0.000000000 -0.009728154

外：

Sample    Gene            Pre.BI            Pre.BI   Immature     Naive         Memory       Plasmacell
MMRF_1021 ENSG00000007062 0.000342301       0        0            0             0             0
MMRF_1021 ENSG00000012124 0                 0        0            0             0            -0.001307111
MMRF_1024 ENSG00000007062 8.015672e-05      0        0            0             0             0
MMRF_1024 ENSG00000012124 0                 0        0            0             0            -0.002693559
.....

dput df：

structure(list(MMRF_1021 = c(0.0537454710193116, 0.134363677548279
), MMRF_1024 = c(0.0125855939107651, 0.276882875966623), MMRF_1029 = c(0, 
0.278044754955015), MMRF_1030 = c(1.29850876031527, 0.715843203834688
), MMRF_1031 = c(0.37618693249153, 0.032711952160723)), row.names = c("ENSG00000007062", 
"ENSG00000012124"), class = "data.frame")

推杆权重：

structure(list(Pre.BI = c(0.006368928, 0), Pre.BII = c(0, 0), 
Immature = c(0, 0), Naive = c(0, 0), Memory = c(0, 0), Plasmacell = c(0, 
-0.009728154)), row.names = c("ENSG00000007062", "ENSG00000012124"
), class = "data.frame")

Answer 1

我想你可能正在寻找这个：

library(tidyverse)
joinedDataframe <- df %>%
rownames_to_column("gene") %>%
gather("sample", "value", -gene) %>%
left_join(weights %>%
rownames_to_column("gene")
, by = "gene")
joinedDataframe %>%
mutate(Pre.BI = Pre.BI * value
, Pre.BII = Pre.BII * value
, Immature = Immature * value
, Naive = Naive * value
, Memory = Memory * value
, Plasmacell = Plasmacell * value) %>%
select(-value)
gene    sample       Pre.BI Pre.BII Immature Naive Memory    Plasmacell
1  ENSG00000007062 MMRF_1021 3.423010e-04       0        0     0      0  0.0000000000
2  ENSG00000012124 MMRF_1021 0.000000e+00       0        0     0      0 -0.0013071105
3  ENSG00000007062 MMRF_1024 8.015674e-05       0        0     0      0  0.0000000000
4  ENSG00000012124 MMRF_1024 0.000000e+00       0        0     0      0 -0.0026935593
5  ENSG00000007062 MMRF_1029 0.000000e+00       0        0     0      0  0.0000000000
6  ENSG00000012124 MMRF_1029 0.000000e+00       0        0     0      0 -0.0027048622
7  ENSG00000007062 MMRF_1030 8.270109e-03       0        0     0      0  0.0000000000
8  ENSG00000012124 MMRF_1030 0.000000e+00       0        0     0      0 -0.0069638329
9  ENSG00000007062 MMRF_1031 2.395907e-03       0        0     0      0  0.0000000000
10 ENSG00000012124 MMRF_1031 0.000000e+00       0        0     0      0 -0.0003182269

Answer 2

看到你的预期结果，我认为以下是你所追求的。例如，MMRF_1024 ENSG00000012124的Plasmacell为 -0.002693559 (0.27688288 * -0.009728154(。为了得到这个数字，我将两个数据帧都转换为长格式数据。然后，我加入了他们。此时，您有两列来处理乘法(即gene_value和值(。在此之后，我将数据转换为宽格式数据框。

librrary(tidyverse)
rownames_to_column(df) %>% 
pivot_longer(cols = -rowname, names_to = "gene", values_to = "gene_value") -> temp1
rownames_to_column(weights) %>% 
pivot_longer(cols = -rowname, names_to = "variable", values_to = "value") -> temp2
left_join(temp1, temp2, by = "rowname") %>% 
mutate(answer = gene_value * value) %>% 
pivot_wider(id_cols = rowname:gene, names_from = "variable", values_from = "answer")
rowname         gene         Pre.BI Pre.BII Immature Naive Memory Plasmacell
<chr>           <chr>         <dbl>   <dbl>    <dbl> <dbl>  <dbl>      <dbl>
1 ENSG00000007062 MMRF_1021 0.000342        0        0     0      0   0       
2 ENSG00000007062 MMRF_1024 0.0000802       0        0     0      0   0       
3 ENSG00000007062 MMRF_1029 0               0        0     0      0   0       
4 ENSG00000007062 MMRF_1030 0.00827         0        0     0      0   0       
5 ENSG00000007062 MMRF_1031 0.00240         0        0     0      0   0       
6 ENSG00000012124 MMRF_1021 0               0        0     0      0  -0.00131 
7 ENSG00000012124 MMRF_1024 0               0        0     0      0  -0.00269 
8 ENSG00000012124 MMRF_1029 0               0        0     0      0  -0.00270 
9 ENSG00000012124 MMRF_1030 0               0        0     0      0  -0.00696 
10 ENSG00000012124 MMRF_1031 0               0        0     0      0  -0.000318

Answer 3

下面是一个基本的 R 解决方案

dfout <- do.call(rbind,
c(make.row.names = F,
lapply(seq(ncol(df)), 
function(k) cbind(Gene = rownames(df[k]), 
Sample = names(df[k]), 
df[,k]*weights[match(rownames(weights),rownames(df)),]))))

这样

> dfout
Gene    Sample       Pre.BI Pre.BII Immature Naive Memory    Plasmacell
1  ENSG00000007062 MMRF_1021 3.423010e-04       0        0     0      0  0.0000000000
2  ENSG00000012124 MMRF_1021 0.000000e+00       0        0     0      0 -0.0013071105
3  ENSG00000007062 MMRF_1024 8.015674e-05       0        0     0      0  0.0000000000
4  ENSG00000012124 MMRF_1024 0.000000e+00       0        0     0      0 -0.0026935593
5  ENSG00000007062 MMRF_1029 0.000000e+00       0        0     0      0  0.0000000000
6  ENSG00000012124 MMRF_1029 0.000000e+00       0        0     0      0 -0.0027048622
7  ENSG00000007062 MMRF_1030 8.270109e-03       0        0     0      0  0.0000000000
8  ENSG00000012124 MMRF_1030 0.000000e+00       0        0     0      0 -0.0069638329
9  ENSG00000007062 MMRF_1031 2.395907e-03       0        0     0      0  0.0000000000
10 ENSG00000012124 MMRF_1031 0.000000e+00       0        0     0      0 -0.0003182269