Entering edit mode
5.5 years ago
au.rinki.bio
▴
20
hello, I had identified differentially expressed genes using Limma for microarray data (cancer). what are the accuracy that these genes truly involved in this disease? there are <1000 genes are reported for particular cancer in different databases. from our analysis, we found few genes those are already reported but why we didn't get those are also reported as biomarkers for this cancer?
Disease association doesn't really imply disease involvement. And you have to be able to reproducibly show that your genes are discriminatory or you multi-gene signature is valid before you get to call them/it a biomarker. Depending on the application (early detection, prognosis, ...), very few biomarkers make it through multiple rounds of validation
I guess, this would be a good exercise to see the data in a different light. Assuming that your data analysis is statistically sound and one or two genes can be experimentally validated, look at the expression patterns of biomarkers in the data. You would get an idea how the biomarkers behave a sound statistically analyzed data, in the context of your lab experiments. You can develop a model based on that.
Please use a more descriptive title for your thread.