普通矩阵转化为稀疏矩阵

之前见到的矩阵都是稠密矩阵：如

myfiles <- list.files(pattern = "*matrices.csv")
myfiles
chow1<-read.csv(myfiles[1],sep=',',header=T,row.names = 1,stringsAsFactors = F)
chow1[1:10,1:5]
#              AAACCTGAGAGTAATC AAACCTGGTAAATGTG AAACGGGAGAAACGAG AAACGGGAGGAGTACC AAACGGGCACCACCAG
#Sox17                        1                0                0                1                0
#Mrpl15                       0                0                0                0                0
#Lypla1                       0                0                0                0                0
#Tcea1                        0                0                0                0                1
#Atp6v1h                      0                0                1                1                0
#Rb1cc1                       0                0                0                0                0
#4732440D04Rik                0                0                0                0                0
#Pcmtd1                       1                0                2                1                1
#Gm26901                      0                0                0                0                0
#Rrs1                         0                0                0                0                0

看看占了多少内存

format(object.size(chow1),units = "Mb")
#[1] "301.3 Mb"

转化为稀疏矩阵

chow1 <- as(as.matrix(chow1), "dgCMatrix")
chow1[1:10,1:5]
#              AAACCTGAGAGTAATC AAACCTGGTAAATGTG AAACGGGAGAAACGAG AAACGGGAGGAGTACC AAACGGGCACCACCAG
#Sox17                        1                .                .                1                .
#Mrpl15                       .                .                .                .                .
#Lypla1                       .                .                .                .                .
#Tcea1                        .                .                .                .                1
#Atp6v1h                      .                .                1                1                .
#Rb1cc1                       .                .                .                .                .
#4732440D04Rik                .                .                .                .                .
#Pcmtd1                       1                .                2                1                1
#Gm26901                      .                .                .                .                .
#Rrs1                         .                .                .                .                .

看看占了多少内存

format(object.size(chow1),units = "Mb")
#[1] "58.8 Mb"

再转化为普通矩阵

chow1 <- as.matrix(chow1)