Entering edit mode
7.8 years ago
goro
•
0
If two diseases share the same pathways. These pathways are an output of enrichment analysis. how can this be explained? Can we assume that the mutations in these pathways are responsible for both of the diseases. Is this assumption correct or the story is something different.
The question is a tiny bit vague, but is a pathway not a permutation of various genes and their expression levels? I'd think any number of diseases could result from a pathway getting out of whack, depending on which gene malfunctions.
I personally know that over-expression and under-expression of genes in the heme biosynthesis pathway (involving 8 principal genes) can cause at least 11 different diseases (that I know of).
Thank you Ram! Let's say we have a list of 100 genes responsible for causing a disease. Another list of 200 genes are responsible for another disease.
Can we conclude that the overlap in the genes between these two lists of genes is responsible for both diseases. what further investigations are needed to check this probable association between the two diseases ?
Someone with experience in pathway enrichment analysis will have to help you with that. I had a thought on the underlying concept and shared that in my post - what you have asked above is beyond me, unfortunately.
I'd be wary of making such conclusions straight away. The data you've could however give you a starting point for more research (in silico, in vitro and in vivo). Have you looked at protein-protein interactions? This papers may be worth a read, High-throughput methods for identification of protein-protein interactions involving short linear motifs and Comparative assessment of large-scale data sets of protein–protein interactions. There is a recent initiative (check this post) interested in integrating biological and chemical data so that it's easier to investigate if diseases sharing the same biological pathways could be treated by the same known drug, so that the latter could be repurposed, for example. You can perhaps have a look at the overlapping genes you've got and see what the Target Validation Platform has to show for them. You can find for example the diseases associated with a gene (e.g. BRAF, find out non-disease related information for BRAF, which genes (or targets) are associated with a disease (e.g. asthma), the drugs available for that disease and its ontology (navigate through the tabs Drugs and Disease Classification for asthma for example).
If you happen to have known lists for both diseases then the overlap is only a list of common genes involved in both. Concluding any more than that is a non sequitur. What more can be concluded will depend entirely on the size of the overlap and what's known regarding how much of each disease can be explained by each list.
Could it not be that your set of mutations influence 'disease A', and 'disease A' influences 'disease B'. You may not be able to assume that the set of mutations influence disease B independently of disease A.