Quality Control of Expression Profiling Data

Mikhail Soloviev, Andrew Milnthorpe

Research output: Contribution to journalArticlepeer-review

Abstract

Expression profiling is a popular tool for studying gene expression levels, but libraries’ origins and data quality are often poorly annotated or contain errors. Experimental techniques, library annotations and analysis algorithms vary between laboratories and may contain errors. Traditional analysis methods, including research into tissuespecific expression, assume expression levels to be correct and libraries to be correctly annotated, which is not always the case. Therefore, tools capable of assessing the quality of multiple types of expression data using the data alone would be invaluable for quality control of that data and elucidation of its suitability for expression analysis. Here we compare and review over 20 methods and focus on a number of key developments in the field. We also highlight the application of recently devised novel quality control methods and show examples of applications of the newly developed quality control expression matrixes (QCEM) to the analysis and quality control of SAGE data. The described example include elucidating the correct tissue identity and show that disease state for expression libraries created using a range of expression profiling methods might be easily elucidated. The described novel quality control methods address key shortcomings of the previously reported tools and provide a universal quality control method for multiple types of expression data.
Original languageEnglish
Article number7
Pages (from-to)176-187
Number of pages12
JournalJournal of Proteomics & Bioinformatics
Volume8
DOIs
Publication statusPublished - 22 Jul 2015

Cite this