Have I used Deseq2 ,on protein array data, correctly? Or should I be using another R package
1
0
Entering edit mode
@reubenmcgregor88-13722
Last seen 3.2 years ago

We had some protein microarrays done I believe using the hu-prot platform. I asked for non-normalised corrected data hoping to run in DeSeq2 (as I have used this for RNAeq data before).

The count data imported as "cts" looks something like this:

  RF0064 RF0065 RF0070
COL4A3BP 2927 4186 5818
IRAK4 9072 63923 5558

 

Patient ID's in columns and proteins in rows, obviously allot more of each. One note here I had to round values up so there where no decimal places for DeSeq

My metadata looks like this in "coldata": 

id case_control days card
RF0064 case 33 carditis
RF0065 control 35 control_NA
RF0070 case 33 non_carditis

 

Again allot more patients.

Here is an example of case vs control analysis I did

dds_case_control <- DESeqDataSetFromMatrix(countData = cts,
                              colData = coldata,
                              design = ~ case_control)

dds_case_control$case_control <- relevel(dds_case_control$case_control, ref="control")


dds_case_control <- DESeq(dds_case_control)

res_case_control <- results(dds_case_control)

summary(res_case_control)

Which all seems to work fine, but I am not an expert at all with this and wondered if there is some ways within Deseq2 to quality control the results? Or if perhaps if I should use a different R package? I do like Deseq2 as I am now familiar with it, but ay advice welcome.

r deseq2 protein microarray • 1.4k views
ADD COMMENT
2
Entering edit mode
@gordon-smyth
Last seen 1 hour ago
WEHI, Melbourne, Australia

You are right to want to quality control.

Microarrays do not produce count data, certainly not negative binomial counts. So it is not meaningful to analyse protein microarrays using DESeq2, DESeq or edgeR.

I am not familiar with the hu-prot platform (by which I think you mean the HuProt platform from CDI), but the limma package has been suggested by others for this sort of data.

I went to the CDI website to find articles using HuProt, and the most recent article did their analysis using limma:

http://www.mcponline.org/content/early/2017/01/13/mcp.M116.063602.abstract

 

ADD COMMENT
0
Entering edit mode

Thank you,

I assumed if I had some kind of data that was equivalent to “counts” I.e background subtracted and corrected data, so essentially have protein abundance data, that the analysis would be equivalent. Showing my ignorance I guess.

ADD REPLY
1
Entering edit mode

No, a fluescent intensity is not a count, and becomes even less like a count after background subtraction.

ADD REPLY

Login before adding your answer.

Traffic: 749 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6