Frozen robust multiarray analysis (fRMA)

Abstract

Robust multiarray analysis (RMA) is the most widely used preprocessing algorithm for Affymetrix and Nimblegen gene expression microarrays. RMA performs background correction, normalization, and summarization in a modular way. The last 2 steps require multiple arrays to be analyzed simultaneously. The ability to borrow information across samples provides RMA various advantages. For example, the summarization step fits a parametric model that accounts for probe effects, assumed to be fixed across arrays, and improves outlier detection. Residuals, obtained from the fitted model, permit the creation of useful quality metrics. However, the dependence on multiple arrays has 2 drawbacks: (1) RMA cannot be used in clinical settings where samples must be processed individually or in small batches and (2) data sets preprocessed separately are not comparable. We propose a preprocessing algorithm, frozen RMA (fRMA), which allows one to analyze microarrays individually or in small batches and then combine the data for analysis. This is accomplished by utilizing information from the large publicly available microarray databases. In particular, estimates of probe-specific effects and variances are precomputed and frozen. Then, with new data sets, these are used in concert with information from the new arrays to normalize and summarize the data. We find that fRMA is comparable to RMA when the data are analyzed as a single batch and outperforms RMA when analyzing multiple batches. The methods described here are implemented in the R package fRMA and are currently available for download from the software section of http://rafalab.jhsph.edu.

The Pluripotency Factor NANOG Binds to GLI Proteins and Represses Hedgehog-mediated Transcription

“…Analysis of Nanog and Gli Co-expression-We searched Barcode-annotated samples (26,27) for biological contexts associated with three expression patterns: 1) high expression in Gli1 (frozenrobustmultiarrayanalysis-normalizedandBarcode-standardized Gli1 expression Ն5) and high expression in Nanog (Nanog expression Ն10), 2) high expression in Gli1 (Gli1 expression Ն5) and medium expression in Nanog (1 Յ Nanog expression Ͻ 10), and 3) high expression in Gli1 (Gli1 expression Ն5) and low expression in Nanog (Nanog expression Ͻ1). In Barcode, samples were processed to facilitate cross-data set comparisons by minimizing unwanted variation such as laboratory or batch effects.…”

“…These samples were compiled by the Barcode project and normalized using frozen robust multiarray analysis to ensure that expression levels of each gene can be meaningfully compared across these heterogenous samples (26,27). The biological context of each sample was annotated and curated by Barcode.…”

Knowledge Discovery from Complex High Dimensional Data

“…Diseases can be considered as abnormal 2 The CEL files from the GEO were normalized and summarized for transcripts using the frozen RMA algorithm [55]. Then only the verified (grade A) genes were chosen for further analysis according to the NetAffx probeset annotation v33.1 of Affymetrix (n = 20492 afterwards).…”

Setting Up the Nonlinear Model of Experimental Data From DNA Microarrays

“…Дальнейшее развитие данный подход получил в рамках метода frozen RMA (fRMA) в работе [9], где коэффициенты сродства линейной модели были оценены по выборке из 850 микрочипов. В следующей работе [10] авторами был представлен пакет frmaTools, позволяющий провести предварительную настройку модели на собственноручно подобранной выборке микрочипов.…”