Skip to content

sig_testmlr_tool

Apply given model to predict genes using multiple linear regression.

Synopsis

sig_testmlr_tool [--ds DS] [--model MODEL] [--minval MINVAL] [--maxval MAXVAL] [--use_gctx USE_GCTX] [--xform XFORM]

Arguments

--help, -h : Show this help message and exit

--help_markdown : Print help in Markdown format

--undef_action UNDEF_ACTION : Action to take if an undefined argument is encountered. Default is error. Options are {error|warn|ignore}

-o, --out OUT : Output path

--runtests : Execute functional and unit tests. Default is 0

--rundemo : Run the tool with sample inputs. Default is 0

--rpt RPT : Report folder prefix

--create_subdir CREATE_SUBDIR : Create subfolder within out for saving results. Default is 1

--verbose VERBOSE : Print debugging messages. Default is 1

--encode_url ENCODE_URL : Encode all input URLs. Default is 0

--config CONFIG : Argument configuration file

--ds DS : GCT or GCTX, Input dataset (genes x samples), minimally containing all landmarks genes specified in the model. Non-landmark genes will be ignored

--model MODEL : MLR model output from SIG_TRAINMLR_TOOL to utilize to infer dependent genes

--minval MINVAL : Minimum value for output dataset. Lower values will be replaced with this value. Default is 0

--maxval MAXVAL : Maximum value for output dataset. Greater values will be replaced with this value. Default is 15

--use_gctx USE_GCTX : Outputs text GCT files if false. Default is 1

--xform XFORM : String. Specify any transformations to apply to data prior to training. Default is none. Options are {none|log2|abs|pow2|zscore}

Description

The sig_testmlr_tool takes as input a test dataset (genes x samples) minimally containing all landmark genes and an MLR model (output by SIG_TRAINMLR_TOOL) and output inferred values for dependent genes by applying the model. Note that the values in the input dataset must match the data used for training. The standard L1000 analytical workflow employs normalized log2 transformed expression as input to the model. The --xform argument can be optionally specified to transform the input data. The output is saved as a GCTX file with the inferred dependent genes appended to the original landmark data.

Examples

  • Apply model to landmark data.

sig_testmlr_tool --ds 'test_data.gctx' --model '/filepath/model.gctx'

  • Apply model to landmark data retricting the minimum value in the output to 2.

sig_testmlr_tool --ds 'test_data.gctx' --model '/filepath/model.gctx' --minval 2