R-数字矩阵

Question

I am new to R and I am trying to convert a dataframe to a numeric matrix using the below code

expData <- read.table("GSM469176.txt",header = F)
expVec <- as.numeric(as.matrix(exp_data))

When I use as.matrix, without as.numeric, it returns some numbers (as below)

0.083531    0.083496    0.083464    0.083435    0.083406    0.083377    0.083348"                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
[9975] "-0.00285  -0.0028274  -0.0028046  -0.0027814  -0.0027574  -0.0027319  -0.0027042  

but when I put in the as.numeric, they are all converted to "NA"

I apologize if someone has asked this question before but I can't find a post that solves my problem. Thanks in advance

Answer 1

You have 2 issues. First, if you examine the structure of the data frame, you'll note that the first column is characters:

head(expData)[, 1:4]

            V1         V2         V3         V4
1 YAL002W(cer) 6.1497e-02 6.2814e-02 6.4130e-02
2 YAL002W(par) 7.1352e-02 7.3262e-02 7.5171e-02
3 YAL003W(cer) 2.2428e-02 3.8252e-02 5.4078e-02
4 YAL003W(par) 2.6548e-02 3.6747e-02 4.6947e-02
5 YAL005C(cer) 2.4023e-05 2.3243e-05 2.2462e-05
6 YAL005C(par) 2.0252e-02 2.0346e-02 2.0440e-02

Therefore, trying to convert the complete data frame to numeric will not work as expected.

Second, you are running as.numeric() after as.matrix(), which is converting the matrix to a vector:

x <- as.numeric(as.matrix(expData))
# Warning message:
# NAs introduced by coercion 
class(x)
[1] "numeric"
dim(x)
# NULL not a matrix
length(x)
# [1] 14261302

I suggest you try this:

rownames(expData) <- expData$V1
expData$V1 <- NULL
expData <- as.matrix(expData)
dim(expData)
# [1] 7502 1900
class(expData[, 1])
# [1] "numeric"

Answer 2

You get the NA's when R doesn't know how to convert something to a number.

Specifically, the quotation mark in your output tells me that you have one (several) LNG string of numbers. To see why this is bad, try: as.nmeric("-0.00285 -0.0028274")

I don't know what your raw data is like, but as @alexwhan mentioned, the culprit is probably in your call to read.table

To fix it, try explicitly setting the sep argument (ie, next to where you have header)

I would suggest opening up the raw file in a simple text editor (TextEdit.app or notepad, not Word) and seeing how they are separated. M guess is

   ..., sep="\t" 

should do the trick.