ga RAD::DataLoad::Plugin::LowessNormalizer [options] --cfg_file cfg_file --qid1 quantification_id1 --qid2 quantification_id2


This plug-in performs lowess normalization on the 2-channel cDNA micorarray data using the global or print-tip lowess method implemented by Gregory Grant at CBIL, University of Pennsylvania. The inputs are the quantification_ids of the two related channels. The processed data will be automatically loaded into ProcessIO and ProcessResult tables. Proper entries will also been inserted into ProcessInvocation, ProcessInvocationParam, and ProcessInvQuantification tables.


--cfg_file config_file [require the absolute pathname] The configuration file.

--qid1 NUM Quantification_id for the first channel, which will be used as the numerator in lowess normalization.

--qid2 NUM Quantification_id for the second channel, which will be used as the denominator in lowess normalization.


--debug Debugging output.

--help Get usage; same as usage.

--verbose Lots of output.

--commit Commit the data to database.

--fg mean|median Use mean or median as the foreground signal intensity. [Default = median]

--bg morph|mean|median Use morph, mean, median as the background intensity. [Default = median]

--bg_correction If specified, background correction will be done, i.e. the background intensity will be substracted from the foreground signal intensity; otherwise, no background correction is done by default.

--print_tip_lowess If specified, the print-tip lowess will be done; otherwise, the global lowess is done by default.

--smoothing_parameter NUM The smoother span. This gives the proportion of points in the plot which influence the smooth at each value. Larger values give more smoothness. [Default = 0.3]

--maxiter NUM The number of robustifying interations which should be performed. [Default = 3]

--desc STRING Free text describing the specific process invocation, which will be stored in ProcessInvocation table.

--user STRING The user name, used to set value for row_user_id. The user must already be in Core::UserInfo table. [Default: from $HOME/ file]

--group STRING The group name, used to set value for row_group_id. The group must already be in Core::GroupInfo table. [Default: from $HOME/ file]

--project STRING The project name, used to set value for row_project_id. The project must already be in Core::ProjectInfo table. [Default: from $HOME/ file]


Before one can run this plug-in, the Perl objects that will be loaded need to be created, if these objects do not exist yet.

Make sure the file of the user contains the correct login name [RADrw]. Also, if the group and project differ from the default values in, please specify the proper group and project name on the command line using --group and --project options respectively.


It should be a tab-delimited text file with 2 column: Name and Value. The order and case of these columns are not important, but it is recommended to follow the template for the sake of consistency. See the sample config file sample_LowessNormalizer.cfg in the $PROJECT_HOME/RAD/config/ directory.

Comments line start with #.

Each (non-comment) line should contain exactly only one tab.

The required parameters are as follows:

element_result_table The specific view of the ElementResultImp to which the element result data after quantification are loaded.

flagN The name of the Nth flag in element_result_table which will be used to exclude elements whose flag value equals that specifed by flag_valueN.

flag_valueN The value specified for the flagN.


Written by Hongxian He.


Copyright Hongxian He, Trustees of University of Pennsylvania 2003.