Skip to main content

Table 2 Double-bias-removal location normalization techniques used in this study. These strategies remove both spatial- and intensity effect either in a single step (IG SG LOESS) or in two steps (the remaining thirteen approaches) by combining methods listed in Table 1.

From: Evaluation of normalization methods for cDNA microarray data by k-NN classification

Name Description: Method/Effect/Level
IG SG LOESS* Joint Intensity/Global & Spatial/Global M l = M - lowess(A, rloc, cloc)
IG LOESS-SL LOESS Step 1: IG LOESS/Intensity/Global lowess
Step 2: SL LOESS/Spatial/Local lowess
IL LOESS-SL LOESS Step 1: IL LOESS/Intensity/Local lowess
Step 2: SL LOESS/Spatial/Local lowess
IG LOESS-SL FILTERW3 Step 1: IG LOESS/Intensity/Global lowess
Step 2: SL FILTERW3/Spatial/Local median filter
IG LOESS-SL FILTERW7 Step 1: IG LOESS/Intensity/Global lowess
Step 2: SL FILTERW7/Spatial/Local median filter
IST SPLINE-SL LOESS Step 1: IST SPLINE/Intensity/Global spline
Step 2: SL LOESS/Spatial/Local lowess
IST SPLINE-SL FILTERW3 Step 1: IST SPLINE/Intensity/ Global spline
Step 2: SL FILTERW3/Spatial/Local median filter
IST SPLINE-SL FILTERW7 Step 1: IST SPLINE/Intensity/Global spline
Step 2: SL FILTERW7/Spatial/Local median filter
Q SPLINEG-SL LOESS Step 1: Q SPLINEG/Intensity/Global qspline
Step 2: SL LOESS/Spatial/Local lowess
Q SPLINEG-SL FILTERW3 Step 1: Q SPLINEG/Intensity/Global qspline
Step 2: SL FILTERW3/Spatial/Local median filter
Q SPLINEG-SL FILTERW7 Step 1: Q SPLINEG/Intensity/Global qspline
Step 2: SL FILTERW7/Spatial/Local median filter
QS PLINER-SL LOESS Step 1: QS PLINER/Intensity/Global qspline
Step 2: SL LOESS/Spatial/Local lowess
QS PLINER-SL FILTERW3 Step 1: QS PLINER/Intensity/Global qspline
Step 2: SL FILTERW3/Spatial/Local median filter
QS PLINER-SL FILTERW7 Step 1: QS PLINER/Intensity/Global qspline
Step 2: SL FILTERW7/Spatial/Local median filter
  1. * IG SG LOESS was implemented in the following package/function: MAANOVA R package/smooth (method="rlowess", f = 0.4, degree = 2). Elsewhere, IG SG LOESS is known as joint loess [21]. lowess(A, rloc, cloc): lowess curve fitted as a function of average log intensity (A), row location (rloc), and column location (cloc) of spots on a microarray.