How can I be sure that I've made a good GO analysis?
1
1
Entering edit mode
9.3 years ago
tiago211287 ★ 1.4k

How can I be sure that I've made a good GO analysis?

I've studing about GO analysis has been 1 month already. I try to use david, Gorilla, Cytoscape(BinGO, clueGO) until now, but I still can not grasp what is really need to say:

"Thats is a good GO analysis and now I am done of it."

Can some more experienced guy show some enlightment in my path?

RNA-Seq Ensembl Gene Onthology • 2.5k views
ADD COMMENT
0
Entering edit mode

The next step would be to consider the methods used by each to calculate enrichment and decide of what you want to make of the individual runs. Compare the similarities and differences in the findings, spot check against data. If you're only using GO (and not other ontologies/databases), you could consider some kind of consensus or simply reporting results from multiple methods, but this likely isn't a benchmarking study.

Make sure that the programs are all using the latest ontology and GOA versions, some don't get updated frequently. In general newer methods may perform better since they're designed to address some of the drawbacks of older methods.

Here's a good read: http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1002375

I'd go with the method that gave you the results that you feel are the best and can readily defend. This is a place where you'll have to scratch your head and use your best judgement.

ADD REPLY
0
Entering edit mode

Thank you very much for your help.

ADD REPLY
0
Entering edit mode
9.3 years ago
andrew ▴ 560

Another thing to consider are the correction factors. Many of the tools out there only allow for correction using FDR or Bonferoni which do not consider the hierarchical structures/relationships of GO terms. Two other methods to consider using are Elim and Weight pruning. Here is a paper describing the methods.

Our tool, iPathwayGuide, has these correction factors built in and is very easy to use. Anyone can try it for free at www.AdvaitaBio.com. Simply click Register to get started. It costs nothing to upload your data and get results. Results are available for 72-hours. or you can pay to keep them forever.

You can get more info here.

ADD COMMENT

Login before adding your answer.

Traffic: 1059 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6