Hello.
I have been working on a transcriptomics data from NCBI's GEO for a while now. However, I have recently been made aware of this phenomenon [see image attached]:
- When plotting the probes (y-axis) against the subjects (x-axis), the heatmap generated shows a very large area (around 10,000 probes) with an intensity lower than their vicinity, both for control and for diseases individuals;
- This effect is also (more visible) on the other image, that shows that for around 10,000 probes (x-axis), the intensity (y-axis) is lower than the average intensity.
My question is: do you have any idea what this could be due to? Is this "area" of lower intensity a common thing in microarray analysis? Should I exclude this area from the analysis?
Thank you in advance
see this link
https://www.ncbi.nlm.nih.gov › NCBI ›
https://bmccancer.biomedcentral.com/article
It seems you did not scale your data when you plotted the heatmap, and you did not use hierarchical clustering to show the relationship between your samples. So your heatmap is full of low absolute expression probes, but not the relative expression after scaling.
The low expression of microarray data is not common, but for RNA-seq is very often. Your data should be microarray data, so the best way is to obtain CEL file or other raw data from the author to repeat the normalization process by yourself.
There are several ways to "trim" your data. You can use boxplots and/or 2-dimensional principal component analysis (PCA) to identify outliers samples, then normalizing and repeating the process, until you can distinguish samples from different groups.
Other approach is to use NUSE plots, density plots, etc. You can find plenty information about preprocessing and data filtering in many microarray data analysis books. Particularly i like this one "Statistics and Data Analysis for Microarrays Using R and Bioconductor". You can find it easily on the internet.
I suggest to figure out where these probes map to. Are these all non-genic or control probes and thus can be discarded? Maybe not, because 10.000 probes are too many. Anyway, it is strange that the effect is shared among all samples.
Which microarray chip is this?
Thus, first thing: what do these probes target?
It is a Illumina HumanHT-12 V3.0 expression beadchip. Their manifest files states that there are some control probes (
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics