What could explain the difference in RNA-seq and Microarray expression levels?

Question

[N]

What are some possible reasons why some trends observed in Microarray expression levels is not observed in RNA-seq. Example the difference between 2 cell types for a gene of interest show major differences when microarray is used with one cell type clearly having higher expression but rna-seq for the same cell types shows equal expression levels for the gene of interest?

What are possible explanations for the discrepancy between such results ? Shouldn't they have relatively identical trends otherwise gene expression in one of them is unreliable?

Answer 1

[Mensur Dlakic]

In a perfect world, women would be doing movie trailer voice-overs just the same as men; Biostar users would search for existing answers to their questions before posting them; only qualified people would be presidents; and yes, experimental results from different methods would be well-correlated.

There are many possible reasons for this discrepancy, starting with human error in preparing one or both samples. Or the human error in data analysis parts of the experiment. Or the microarray is not well designed and doesn't capture the expression well.

RNA-seq should be unbiased because it is not probe-dependent, and also has a higher dynamic range as it is more unlikely to reach saturation. RNA-seq also can identify transcripts that were unknown before the experiment began, including alternatively spliced isoforms. All things being equal in sample prep and analysis, RNA-seq data are more likely to be reliable.