Dear biostars,
what's the best way to enrich a set of marker genes in microarray analysis ? Imagine you have 30 marker gene (selected set of gene_id that are known ) and you want to enrich this set of genes and pull out more informative genes from your microarray to analysis your microarray based on expression value of the new set, so what you will do?
If you have a set of marker genes, presumably defined as exhibiting a given pattern in a given data set, and you'd like to find more of those kinds of genes, one approach would be to simply take your marker genes and fish for more genes showing the same pattern. The genefilter package in R has a function for doing this called genefinder(). Given a data set (data from several microarrays), and a pattern you'd like to find, it will pick out the closest genes to your query pattern. Thus for any one of your marker genes, you could find genes with similar patterns. Of course, the idea is simple, and there are plenty of ways to do a similar thing without using R.
You could also try clustering your data and examine the clusters that your marker genes co-cluster with. In both of these cases, you would be enriching your starting set, by adding more genes with similar patterns. As to whether they are "more informative", that's another issue, presumably dependent on what you mean by "more informative". Perhaps some of these genes would add more weight to some biological process not easily identified from your original set.
finding genes with similar pattern of expression in different experiments is a second priority in my analysis, I would like to find new members for my list (enrichment/enlargement) with helps of previous studies, I think literature will be very helpful in finding new informative genes, since they will suggest experimentally-proven genes which have a relationship with marker genes... any suggestion ?
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
You'll need to be somewhat more specific here. What is your experimental design (number of replicates, time points, etc.)? What is the biological question you are trying to answer? The question as you have stated it above is based on some biological hypothesis, but we do not know what that hypothesis is.
7 time-point and 3 rep, I'm trying to enrich/enlarge a set of genes with helps of what others find previously in their studies (GO) , finding new members for a pathway (neuronal cell differentiation ) and then do your analysis to see the relationship between genes in new set.