MicroRNA-microarray data analysis in the precence of FFPE storage time effects
2010 (English)Manuscript (preprint) (Other academic)
Background: The standard method for preserving patient samples for diagnostic purposes is fixation in formalin followed by embedding in paraffin (FFPE). The use of FFPE blocks makes it possible to include a large number of patients in the experimental studies since millions of FFPE blocks are stored around the world. However, FFPE storage can cause degradation and modifications of nucleic acids. In order to draw reliable biological conclusions it is therefore important to know what effect FFPE-storage have on the tissues and to have procedures that normalize this effect. In this paper, we study the effect that FFPE-storage has on microRNA-microarray data from tongue-cancer patients and propose a novel procedure for normalizing the bias introduced by FFPE-storage.
Results: MicroRNA-microarray data from 21 tongue-cancer patients and 8 control patients were used. The samples were stored in FFPE blocks and had been in storage for up to 11 years. The data contained a large amount of biological relevant variation, yet the largest variation was due to the samples storage times. The storage effect was shown to be significant and some results suggested that it may be causal. Moreover, the microRNAs were unequally affected by storage and this could partially be explained by sequence characteristics. The novel normalization procedure was shown to have a large impact in the analysis ability to identify differentially expressed microRNAs between young and old cancer patients as well as between cancer and control patients. The p-values for the top microRNAs candidates were much lower for the proposed novel normalization compared to a standard normalization procedure which suggested that the novel normalization made the analysis more efficient.
Conclusions: MicroRNA-microarray data can be seriously affected by FFPE-storage and the introduced variation cannot be removed by standard normalizations. The proposed normalization removes the bias introduced by FFPE-storage and gives higher sensitivity than the standard normalization.
Place, publisher, year, edition, pages
microRNA, microarray, FFPE, storage time effects, normalization
Probability Theory and Statistics
Research subject Statistics
IdentifiersURN: urn:nbn:se:umu:diva-33355OAI: oai:DiVA.org:umu-33355DiVA: diva2:311613