I want to infer gene regulatory interactions from microarray gene expression data. I have calculated correlation matrix (Pearsons correlation coefficient) on the differentially expressed genes. Further I am unable to find out regulatory interactions that which technique can be used to infer activation and inhibition interactions. Please guide me
If you have enough samples e.g. 200 you can use ARACNE to infer a regulatory network and render it in cytoscape. You wouldn't use the differentially expressed genes to do that.
hi,
you should take a look to the "NetworkInference" biocView of the Bioconductor software that will list for you all the Bioconductor software packages that categorize themselves as suitable for inferring networks from data. You can find this by going to http://www.bioconductor.org, follow the "Install" tab, then the link called "Software" on the right and then on the left under "biocViews" follow "Software" -> "BiologicalQuestion" -> "NetworkInference". The direct link is:
http://www.bioconductor.org/packages/release/BiocViews.html#___NetworkInference
Among the available packages I can advice you to use qpgraph since I'm its maintainer :) ln its documentation the vignettes entitled "BasicUsersGuide.pdf" and "Reverse-engineer transcriptional regulatory networks using qpgraph" are good starting points.
cheers,
robert.
Can you please guide me how to obtain real interactions amongst genes from microarray data. like correlation returns diagonal and symmetrical matrix. I need such a correlation measure in which gene A to gene A interaction is not always equal to 1, also interaction value of gene A to Gene B must be different from interaction value of Gene B to Gene A.
regards
The kind of associations you are referring to, self-loops and directed associations, can in general be only reliably inferred from time-course dynamic data. Unfortunately, qpgraph does not provide methods for such type of data. In the biocViews of NetworkInference that I was referring to in my first answer above of two months ago, I see there are two packages that seem to work with dynamic data, GRENITS and rHVDM. Take a look at them and see whether they work for you. I cannot give you further advice.
cheers,
robert.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
I have read qpgraph. The problem is that it is using data with some ionteractiondb as an input but i dnt have any interaction db. i just have 24 samples og huma gene data and i want to make gene regulatory network by showing gene gene interaction. Please guide
hi,
qpgraph takes an expression data set as input in the form of either an 'ExpressionSet' object, or a 'matrix' object or a 'data.frame' object.
from your comment, you seem to be lost at how to import the expression data into R. i recommend you to have a careful read at introductory materials for R, the internet is plenty of them, google them! and at the Bioconductor workflow for reading and analyzing microarray data, which you can find at http://www.bioconductor.org/help/workflows/arrays
cheers,
robert.