Question: normalization in cross meta-analysis
0
gravatar for Morteza Hadizadeh
4 months ago by
Physiology Research Center, Institute of Basic and Clinical Physiology Sciences, Kerman University of Medical Sciences, Kerman, Iran
Morteza Hadizadeh20 wrote:

Dear all,

I am trying to perform a meta-analysis using microarray data of 6 datasets with different platforms (Illumina HumanHT-12 V4.0 expression beadchip, Affymetrix Human Exon 1.0 ST Array, Affymetrix Human Gene 1.0 ST Array).

My question is about the first step of meta-analysis which is data normalization (before merging the data and remove the batch effect), may I normalize all of the dataset simultaneous (for example using the LIMMA package)? Or must be normalized them separately? Which way is correct and the best way?

Thank you in advance.

R • 194 views
ADD COMMENTlink modified 4 months ago by JC9.4k • written 4 months ago by Morteza Hadizadeh20
1
gravatar for JC
4 months ago by
JC9.4k
Mexico
JC9.4k wrote:

It is never recommended to merge raw data from different technologies, so it's better to normalize/analyze the data separately and then merge your results.

ADD COMMENTlink written 4 months ago by JC9.4k

Thank you for your answer, but in some studies (below links) different microarray platforms have been integrated, the created batch effects after merging data are removed (for example by ComBat function in sva package).

https://www.biorxiv.org/content/10.1101/059600v1.full#disqus_thread

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5697344/

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0194844

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996376/

ADD REPLYlink modified 4 months ago • written 4 months ago by Morteza Hadizadeh20
  1. Normalization is before merge at probeset level https://www.biorxiv.org/content/biorxiv/early/2016/06/18/059600/F1.medium.gif
  2. "The first step in the methodology for microarray data is to put together all the selected series, independently of their technology (Affimetrix or Illumina). Consequently, a quality analysis assessment was performed across the series, in order to detect and consequently remove any possible outlier" -> so they are normalizing before merge the data
  3. The preprocessing is before merge at the gene level https://journals.plos.org/plosone/article/figure/image?size=inline&id=info:doi/10.1371/journal.pone.0194844.g001
  4. "One major limitation of existing cross-platform normalization is that they require that every treatment group or sample type be represented on each platform to allow differentiation of treatment effects from platform effects. Furthermore, cross-platform normalization methods do not guarantee elimination of laboratory or batch effects across experiments"
ADD REPLYlink modified 4 months ago • written 4 months ago by JC9.4k
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: 1673 users visited in the last hour