My favourite gene is expressed at a very low level; will it be correlated with other (random?) unexpressed genes?
Any gene with expression below 20 units on all arrays has been deleted from the dataset, so you won’t get a set of random genes showing correlation only because they’re not expressed. The other genes will have significant expression on at least some arrays (but very low expression on other arrays potentially) and these are all included in the database for the correlation calculations. We have used ACT to analyse some genes which are expressed at very low levels on almost all arrays, only showing significant expression in a small number of samples analysing development of a single tissue (based on results using tools at the NASC and Genevestigator sites). The most highly-ranked genes show very high r-values (>0.95) but with over-representation of annotations from a single area of cell biology giving confidence that the result is meaningful.
Related Questions
- What are theses alternate forms of a gene called? Genes located on a homologous chromosomes may have alternate forms that control different forms of a trait?
- How can lethal genes remain in the gene pool over time instead of disappearing overtime?
- Can the list of positively or negatively correlated genes be saved as a gene list?