Question: Is this a valid analysis?
1
gravatar for randalljellis
2.8 years ago by
randalljellis70 wrote:

Hello,

I analyzed a few GEO data sets that were on different microarray platforms using the RankProd R library.

However, since the data sets were on different microarray platforms, they have different probe sets and hence there were tons of NAs in the all-inclusive expression set. The candidates I ended up taking from this analysis actually ended up working, but I'm wondering if this analysis is defensible.

Thank you in advance for any insight :)

ADD COMMENTlink written 2.8 years ago by randalljellis70
1

Give a try with R package called "virtualarray". If you get similar results, then your approach should be defensible.

ADD REPLYlink written 2.8 years ago by EVR540

Thank you, I will try this.

ADD REPLYlink written 2.8 years ago by randalljellis70

virtualArray seems to not install on R 3.3.1 and is not on GitHub :( do you have any suggestions please?

ADD REPLYlink written 2.8 years ago by randalljellis70
1

You can always shuffle the values and see if you get rubbish or your candidates.

ADD REPLYlink written 2.8 years ago by Asaf6.1k

Thank you, I will try this.

ADD REPLYlink written 2.8 years ago by randalljellis70

I'm not sure what you mean by "shuffle the values," would you mind elaborating? Do you mean mixing up the order of the rows in the expression set?

ADD REPLYlink written 2.8 years ago by randalljellis70

I meant shuffling the values (not NAs) between genes

ADD REPLYlink written 2.8 years ago by Asaf6.1k

I'm sorry, but could you explain this further? Wouldn't doing this totally fudge the whole analysis by changing the entries?

ADD REPLYlink written 2.8 years ago by randalljellis70

Your concerns are that the difference in platforms gave you the set of genes, not their differential expression. You can test this by comparing the results you got with results from fictional data. If you shuffle the values inside each platform and repeat the test you'll expect to find the same results only if your initial results were wrong. If the genes you got are true DE, you would expect them to disappear on shuffled data.

ADD REPLYlink written 2.8 years ago by Asaf6.1k

ahhh, I see. Thank you

ADD REPLYlink written 2.8 years ago by randalljellis70

Hello randalljellis!

It appears that your post has been cross-posted to another site: http://seqanswers.com/forums/showthread.php?p=200788

This is typically not recommended as it runs the risk of annoying people in both communities.

ADD REPLYlink written 2.8 years ago by WouterDeCoster40k
1

When one cannot find any answers in the other forum, Obviously he/she will try to post their questions in the forum where they could find a answer or any guidance. So people can also ignore the "cross-posts"

ADD REPLYlink written 2.8 years ago by EVR540
1

The threads were made just a few hours apart from each other, a bit of patience would have been nice. About cross posting I would like to refer you to How To Ask Good Questions On Technical And Scientific Forums

Avoid cross-posting: Asking the same question in different forums or online communities at the same time is referred to as "cross-posting". Cross-posting is considered rude because you are making two or more communities work for you when only one may be needed. It therefore abuses the time of other users. If some time after posting a question you have not received an answer on a given forum and you feel that posting on another community may be needed, provide a link to the original question within your new post. Once you find or receive a satisfactory answer, be sure to make it available on both forums.

ADD REPLYlink written 2.8 years ago by WouterDeCoster40k

I apologize, I just wanted the greatest possible number of perspectives on this.

ADD REPLYlink written 2.8 years ago by randalljellis70

Thank you for your empathy.

ADD REPLYlink written 2.8 years ago by randalljellis70
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1140 users visited in the last hour