Changeset 879

Timestamp:

Dec 5, 2008, 11:28:12 AM (14 years ago)

Author:

Jari Häkkinen

Message:

Changed documentation to document the Center plug-in

File:

• plugins/base1/se.lu.onk.Center/trunk/README (modified) (3 diffs)

plugins/base1/se.lu.onk.Center/trunk/README

@@ -3 +3 @@
 = About `se.lu.onk.Center` =
 
-The `se.lu.onk.Center` plug-in merge bioassays by grouping them
-according to annotation types or syntactical differences in the
-name. See ``Documentation`` below fo further information about the
-plug-in.
+The `se.lu.onk.Center` plug-in is a normalisation method where
+expression values are centered around the data median or mean. See
+``Documentation`` below for further information about the plug-in.
 
 `se.lu.onk.Center` is free software. See the file license.txt for
@@ -51 +50 @@
 {{{
   # Log in to your BASE as admin
-  # Upload the configuration file `mergebioassay.base`
+  # Upload the configuration file misc/plugin_Transformation_Center.base`
   # Create/update a configuration for the plug-in where you select the
-    plug-in defintion files `mergebioassay.base`. BASE will respond
-    that plug-ins where configured but ...
+    plug-in definition file `plugin_Transformation_Center.base`. BASE
+    will respond that plug-ins where configured but ...
   # Re-configure the plug-in without selecting a file, set the `Plugin
     executables path` and proceed to next window by clicking
@@ -124 +123 @@
 == Documentation ==
 
-This plug-in base-users merge bioassays by grouping them according to
-annotation types or syntactical differences in the name. There are
-four different methods for calculating the merged values. They differ
-in what kind of mean they are using (arithmetic or geometric) and if
-they are merging on ratio or intensity.
+To center your data means that data values are adjusted to reflect
+their variation from some property of the data such as the mean or
+median. The center plug-in allows the user to center the expression
+levels either per gene or per array.
 
-For each group the parents are saved so the link to the raw data is
-preserved. Annotations for the bioassays are saved if sample1 is the
-same for each assay in each group.
+Consider a common experimental design where you are looking at a large
+number of samples all compared to a common reference. For each gene,
+you have a series of ratio values that are relative to the expression
+level of that gene in the reference sample. Since the reference sample
+really has nothing to do with your experiment, you want your analysis
+to be independent of the amount of a gene present in the reference
+sample. This is achieved by center your data on genes. Centering makes
+less sense in experiments where the reference sample is part of the
+experiment.
 
-=== Merge methods ===
+Centering the data for arrays can also be used to remove certain types
+of bias and can be seen as a crude normalization. The results of many
+two-color fluorescent hybridization experiments are not corrected for
+systematic biases in ratios that are the result of differences in RNA
+amounts, labeling efficiency, and image acquisition parameters. Such
+bias have the effect of multiplying ratios for all genes by a fixed
+scalar. Mean or median centering the data in log-space has the effect
+of correcting this bias, although it should be noted that an
+assumption is being made in correcting this bias, which is that the
+average gene in a given experiment is expected to have a ratio of 1.0
+(or log-ratio of 0).
 
-''Geometric mean of ratio''::
-  This method calculates the geometric mean of the ratios in each
-  group per position.
-
-''Arithmetic mean of ratio''::
-  This method calculates the arithmetic mean of the ratios in each
-  group per position.
-
-''Arithmetic mean of intensity''::
-  This method calculates the arithmetic mean of the intensities in
-  each group per position.
-
-''Ratio of ratio''::
-  This method works a little bit different from the others. It
-  requires that each group consists of exactly two bioassays. It will
-  then create a new ratio for each spot with one of the assays over
-  the other, depending on what the user has specified in the ratio
-  parameters.
+In general, it is recommended that median rather than mean centering
+is used since it is more robust against outliers.
 
 === Parameters ===
 
-''Assayname''::
-  This parameter specifies how the name of the assays is
-  constructed. Use %0-%9 as wildcards and then put any character(s)
-  between as delimiters. For example assay name Assay_1_A can be
-  divided into three parts. By setting the parameter to %1_%2_%3 the
-  plugin will read assay name assay_1_A and set %1=Assay, %2=1 and
-  %3=A. Thus, if your assays have names constructed in the same way
-  merge can sort the assays in groups depending on %1, %2 or %3.
+''Center on genes/arrays'' - Whether centering should be done on
+genes, arrays or both. If both is chosen then the centering will first
+be done on genes then on arrays, this is called a cycle.
 
-''Group''::
-  This will specify how the assays will be grouped together. It can
-  either be a %-number used in assayname or an annotation type.
+''Number of centering cycles'' - How many cycles should be done during
+centering. This value is only relevant if the centering should be done
+on both genes and arrays.
 
-  Example 1: Assays; Assay_1_A, Assay_2_A, Assay_1_B and Assay_2_B can
-  be sorted into groups of assays using merge.
-
-  A: Setting the parameter \221Assayname\222 = %1_%2_%3 and parameter
-  \221Group\222 = %2 will sort the assays into two groups.
-{{{
-One group were %2=1  (assays Assay_1_A and Assay_1_B)
-One group were %2=2  (assays Assay_2_A and Assay_2_B)
-}}}
-  B: Setting the parameter \221Assayname\222 = %1_%2_%3 and parameter
-  \221Group\222 = %3 will sort the assays into two groups.
-{{{
-One group were %3=A  (assays Assay_1_A and Assay_2_A)
-One group were %3=B  (assays Assay_1_B and Assay_1_B)
-}}}
-
-''New name''::
-  This text will be used in the name of the merged assays. The name
-  will be this string followed by the groupname.
-
-''Merge method''::
-  Choose one of the methodes described above.
-
-''Ratio''::
-  If merge method `ratio of ratio` is chosen the form of the ratio
-  must be specified. This is done either in the ratio parameter or in
-  ratio annotation type and ratio annotation value. In the ratio
-  parameter you can choose a %-number followed by the two values it
-  can take. The last character in the string must be the delimiter the
-  user has choosen ot use. The other way a user can specify the ratio
-  is by choosing an annotation type in the list and specify the values
-  in the annotation values parameter. This is done by writing the
-  values with the delimiter used last,
-{{{
-%1|vr1|vr2|
-}}}
-  This will take the values from %1 in the assayname and put the assay
-  with value vr1 over the assay with values vr2. Notice that the final
-  character is the delimiter (this can of course be any character).
-
-  Example 2: Assays; Assay_1_A, Assay_2_A, Assay_1_B and Assay_2_B can be
-  sorted into groups of assays using merge.
-
-  A: Setting the parameter \221Assayname\222 = %1_%2_%3 and parameter
-  \221Group\222 = %2 will sort the assays into two groups.
-{{{
-One group were %2=1  (assays Assay_1_A and Assay_1_B)
-One group were %2=2  (assays Assay_2_A and Assay_2_B)
-}}}
-  B: By setting parameter \221Ratio\222 = %3|A|B| ratios of ratios
-  will be created as:
-{{{
-Assay_1_A / Assay_1_B
-Assay_2_A / Assay_2_B
-}}}
-
-=== References ===
-
-[http://mathworld.wolfram.com/ArithmeticMean.html Arithmetic mean] [[br]]
-[http://mathworld.wolfram.com/GeometricMean.html Geometric mean]
+''Centering using median or mean'' - Whether median or mean should be
+used for the centering.