I am working in R and have micorarray expression matrices from Agilent in log10 format. These matrices have the following colums:
Structure: BrightCorner, E1A, Structural, and blank
Probe: (-)3xSLv1, (+)E1A_..., A_..., DarkCorner, DCP_..., ERCC-..., ETG..., and GE_BrightCorner.
And then Gene Symbol.
Plotting the expression of the first 20 subjects (attached boxplots.pnf file) shows that the data is probably not normalized.
Since I am used to working with data that has already been properly processed, I am not sure of how should I process this information to obtain a proper expression matrix. A very naive approach would be to simply ignore any row that wasn't mapped to a gene symbol, collapse expression by gene symbol, anti-log the values, normalize between arrays and calculate log2. But I don't know what kind of information is stored in all the other rows and how I should use it.
Hi Dario,
What is the name of the genome you are working with? Usually Agilent has explanation of what the rows and the columns in an array represent. I suggest you refer to the Feature extraction (the software used to process agilent data) manual or contact the support contact of agilent or the distributor from where you got your data. Based on your explanation, it seems the columns
Structure: BrightCorner, E1A, Structural, and blank
Probe: (-)3xSLv1, (+)E1A_..., A_..., DarkCorner, DCP_..., ERCC-..., ETG..., and GE_BrightCorner are control columns. Please refer to the following document for more details:
https://earray.chem.agilent.com/suredesign/help/Agilent_control_grids.htm
Good luck with your analysis.
Hello Dr. Motwani,
I am working with human genome. Sadly, I am having troubles contacting the team that generated the data and I don't know exactly how they performed the feature extraction. Also, since I am not familiar with the process of going from the lab to a ready-to-analyze expression matrix, I am still not sure whether the control probes you mention should be used to further process/normalize the data or they are just leftovers from feature extraction that no longer serve any purpose and can be discarded.
The link you provided helps a lot, it has also led me to finding out the Feature Extraction software from Agilent and I guess that reading it will give me a much clearer picture of what I am dealing with.
Thanks a lot.
Hello. You can try our software eUTOPIA.
https://scfbm.biomedcentral.com/articles/10.1186/s13029-019-0071-7
- Badges
- Science method