significant differential expressed gene validation
3
0
Entering edit mode
8.6 years ago
zizigolu ★ 4.3k

hi,

I should do some bioinformatics analysis of 8 CEL files derived from heat treatment versus control in Arabidopsis. I am going to perform differential expression analysis then clustering of these genes and enrichment of clusters

for publication how many and what genes I should choose for qPCR validation ???

thank you

R gene • 2.3k views
ADD COMMENT
0
Entering edit mode

You seem to be asking us to do your research for you. How can anyone in the forum possibly know how many genes you would need to pick, let alone which ones?

ADD REPLY
0
Entering edit mode

sorry Ram, I am asking about the criteria for choosing these genes might my computational analysis be publishable

ADD REPLY
2
Entering edit mode

Your experiment deals with heat treatment, correct? Maybe read up on Arabidopsis pathways involved in heat response and target genes from there?

Ideally, research should ask questions and if you find interesting answers, you publish them. You cannot seek to ask only those questions that would lead to a publication (Maybe I'm being naive there). I don't think you can gain experience if you don't dig up the information yourself.

Maybe work on it a bit and start a forum post to ensure you're on the right track? People will be more willing to help you out then.

ADD REPLY
1
Entering edit mode

You need to analyse the data first to see if it does show any differential expression, then decide whether there are any results you want to validate by qPCR. Not all experiments lead to publishable results.

ADD REPLY
0
Entering edit mode

thank you... research should ask questions and if you find interesting answers, you publish them

ADD REPLY
3
Entering edit mode
8.6 years ago
Shyam ▴ 150

Once you analyze the data and identified the DEGs, classify them into different metabolic pathways and gene networks based on the GO, literature and known drought response pathways etc. once you have these you can pick few genes from the from these pathways or networks and validate by qPCR. I would go for the known genes in these from the literature. Then pic any new genes you may identify in your data or any genes with contradictory results for validation. Number depends on the resources your can afford or want to put in this project. The number and type of genes is really a matter of convincing the reviewers and readers and depends on what you really want to achieve from the project. Hope this will help.

ADD COMMENT
3
Entering edit mode
8.6 years ago
Shicheng Guo ★ 9.6k

First, You should Understand why you need to validate the result.

1) The DGE you obtained might be false positive caused by fluorescence in beadchip, you need use other method to validate the result, such as RT-PCR (validation proecess).

2) The DGE what you obtain might be only right in your samples, therefore, you need validate the expression in another dataset (replication process).

3) the DGE might be just correlation or consequence, you might want to find some causal genes which is play important role in this process (Biological interpretation).

4) Finally, you can also use another GEO dataset to validate you result, rather than experiments by yourself.

When you prove these question to the reviewer, you job will be publishable.

ADD COMMENT
0
Entering edit mode

sorry,

I performed my analysis and some genes like AtGolS1 and GLP4 were strongly up and down-regulated respectively between treatment and control and these genes all are related to which i was investigating. now I should valid my analysis by qPCR

is it enough to order primer for five of outstanding up and down-regulating genes????

actually I am not sure how to select my genes

do you have more comments please?

ADD REPLY
2
Entering edit mode

You should really be speaking to your advisor, or seeking an alternative on campus that can guide you.

ADD REPLY
0
Entering edit mode

Read a bit more literature about the concerned genes in the concerned experimental types and species. You are asking too much specific stuffs as per your requirement which is out of ones comprehension and also we should not do it. As Ram pointed, speak to someone more experienced in the lab or your supervisor with the list of genes and see what he has to comment.

ADD REPLY
3
Entering edit mode
8.6 years ago
ivivek_ngs ★ 5.2k

There are few things which you can do, but keeping in mind as Ram said is to dig a bit more into the literature of Arabidopsis sp where you can find publications having heat treatment vs control.

So there are many ways to do it:

1) Perform differential expression and find the top 150 DEGs (assess the significance of those DESs in terms of enrichment)

2) Take from literature of such heat treatment publications for your desired strain and try to see how many of the genes that you have found as DEGs overlap with the DEGs of the publication. If there is a significant amount of hit try to make a hypergeometric estimation of it and see if that overlap is significant or not. Asses the fold change of those hits. Fish out the pathways they are involved and see if these are important pathways or not and then select those genes that come from the top pathways and have recurrent hits in first few top pathways from your overlap list (this might be too stringent but still if you come out with a panel of 20-30 genes might be worth doing

3) Alternatively you can do pathway analysis for total DEGs and find the top significant pathways and check them in line with the publications to see if these are recurrent pathways or not. If so try to find the enrichment of your DEGs in those pathways to find top scoring enriched genes and you can in fact validate those genes

4) I do not usually prefer much GO categories but still you can try. For Biological process and Molecular functions, you can take the top GO terms and see to what over-representations they fall, and see if such terms are reported in your species literature or not for heat treatment designs. If you again try to find enrichment of your DEGs and perform qPCR for the top genes with high enrichment score and significance.

These should work out but then again no estimate of genes can be made directly. You have to find in the literature which are the most frequented GO terms or pathways involved for such work and the genes that concede such and try to see how your DEGs behave among them.

ADD COMMENT
0
Entering edit mode

thank you all for your valuable comments

ADD REPLY
1
Entering edit mode

Accept the answer if its valuable for you so that it can be helpful for others in future.

ADD REPLY

Login before adding your answer.

Traffic: 1865 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