replicate_regression_omics(data_file, foptions_file, base_directory)
Bayesian replicate regression for omics data
FUNCTION ARGUMENTS:
data_file: omics data file (full directory path)
foptions_file: table file containing the options (full directory path)
base_directory: directory name for results (full directory path)
goptions_file (optional): table file containing the options for
single protein graphics (full directory path)
OUTPUT: Data and graphics are written to files
--------------------------------------------------------------------------------
Format of data file (tab-separated text, see examples):
Line 1: Headers
Headers of protein names columns (e.g., !BSUnumber, !BGnumber, !GeneName, !UniprotID),
followed by sample names (as headers of data columns)
Line 2: Time points
first column: !Time data columns: time points (numbers)
Line 3: Replicate numbers
first column: !Replicate data columns: replicate names
Line 4 (optional): !ValueType ('Value', 'Mean', or 'Std')
Further lines: numerical data
--------------------------------------------------------------------------------
Format of options file (tab-separated text, see examples):
Each line contains one attribute:
first column: attribute name
second column: attribute value (string or number)
all further columns are ignored
Lines starting with the '%' character are ignored (can be used for comments)
The attribute 'options_file' allows to declare another options file containing default options
The attribute 'data_file_csv' contains the name of the data file
--------------------------------------------------------------------------------
Attributes in options file (for replicate_regression_omics_analysis)
data_dir directory name for data files
result_dir directory name for result files
graphics_dir directory name for graphics
data_file_csv filename for data file (tsv format, see examples)
data_file_matlab filename for matlab data file (written during the analysis)
options_file filename for default options file (tsv format)
options_out_csv filename for completed options file (tsv format, written during analysis)
translation_table_file filename for ID mapping table (see example)
result_file_matlab filename for
result_file_csv filename for hahne_salt_stress_cytosol_result.tsv
result_file_zip filename for hahne_salt_stress_result.zip
graphics_file file basename for graphics
data_time_unit time unit ('min')
data_scale 'absolute' or 'log2' (also 'ln','log','log10','log2 ratio'; these are all treated like 'log2');
data_min_num_replicates minimal number of valid replicates (genes with less valid replicates are discarded; default 1)
For "absolute" data:
abs_data_adjust_std_upper upper threshold; points above are outliers (increase std dev by factor of 3)
abs_data_adjust_std_lower lower threshold; points below are outliers (increase std dev by factor of 3)
data_std_relative default for relative standard deviation
data_std_minimal minimal standard deviation
For logarithmic data ( 'log2', 'ln','log','log10','log2 ratio')
data_std_log default for standard deviation (on log scale)
log_data_adjust_std_threshold threshold for data values (on chosen log scale) for which std dev is modified
(criterion: | [data value] - [median for this gene & replicate] | > threshold )
for the inserted std dev, see next entry
log_data_adjust_std_factor new std width = factor * absolute deviation from median
data_min_data_points minimal number of data points required in the analysis (default 3)
at least one replicate has to reach this number, points are times t<0 do not count
replicates with less data points are ignored
convert_to_logarithm convert (nonlogarithmic) data to logarithms for replicate regression (Boolean)
log_transformation type of transformation
'arithmetic': data=mean values and plotting on absolute scale
'geometric' : data = median values and plotting on log scale (but data on absolute scale)
ignore_std_deviations Boolean, ignore standard deviations given in data
basis string: type of basis functions
fixed_prior keeping the prior fixed? (Boolean, default 0)
prior_updating number of prior updating iterations (default 10)
updating_factor prior updating factor, default 1.2
updating_factor_final prior updating factor before last regression; default [] (not set)
update_prior_means change parameter means from 0 to posterior means while updating? default 0
t_smooth time constant defining how prior widths depend on the frequency
options_start_value fixed starting value -> to be inserted into options as options.start_value
options_start_at_t Starting time point for changes (after constant behaviour) to be inserted into options as options.start_at_t
options_constant_before_start Boolean (keep curves constant before starting time) to be inserted into options as options.constant_before_start
regression_t_interp time points for regression (optional)
regression_tmin start time for regression (optional)
regression_tmax end time for regression (optional)
crossvalidation run crossvalidation? (Boolean, default 0)
postprocess_normalise Boolean, default 1
graphics_individual file basename (used in script replicate_regression_omics_selected'
graphics_scale default 'log2', 'linear'
graphics_format 'eps', 'png' (for technical reasons, 'eps' needs to be written in single quotes)
convenience_name type of protein names to be used in graphics (default 'SubtiWiki_20090701')
normalise_by_median (Boolean, default TRUE)
mark_outliers_percentage percentage of data points to be marked as outliers based on crossvalidation error
Additional attributes in options file for individual graphics
(function 'replicate_regression_omics_selected')
graphics_scale 'log2','linear'
postprocess_normalise 1
element_id id (or list, selected by |)
element_name name (or list, selected by |)
delimiter_symbol symbol for delimiting list of elements (in element_id, element_name)
title_string title for graphics
x_label x label for graphics
y_label y label for graphics
plot_data produce plot for data (single element)
plot_replicates produce plot with replicates (single element)
plot_regression produce plot for regression curves (single element)
plot_all produce joint plots for all elements