M-CHiPS workspace: What is stored in which variable?

Short overview

Experiment structure

Conditions and measurements
An M-CHiPS workspace contains at least ncon=2 conditions (control plus one further condition, see figure). Variable stain records, which condition comprises how many measurements (sequence as selected in the browser). For single-channel data, measurements resemble hybridizations. For multi-channel data, measurements resemble channels. The total number of measurements is also summarized by nexp. In case of multichannel experiments, there will be a variable multichannel storing the hybridizations affiliated to the measurements (in this case: the channels). Many variables (such as prim, see figure) are of size ngen x nexp, containing conditions and measurements as they were selected in the browser.
Genes
The rows of these matrices resemble the genes. A prerequisite for using M-CHiPS is that genes are spotted in duplicates. Spots are devided into a "primary" and a "secondary" spot set. Please note that one variable only comprises half of the spots while for above example, another variable secu takes the remaining spotset. Number and sequence of rows (as stored in the database, attribute "spotno") do not change throughout analysis. Sorting and filtering of the genes are stored as an index (geneorder) and so are the gene tags (x, xred, xblue, xgreen).

Data
- ngen x nexp data matrices:
  - prim - raw data (primary spot set)
  - secu - raw data (secondary spot set)
  - fitprim - normalized absolute (or estimated) intensities (primary spot set)
  - fitsecu - normalized absolute (or estimated) intensities (secondary spot set)
  - flatprim - normalized (linear) ratios (primary spot set)
  - flatsecu - normalized (linear) ratios (secondary spot set)
  - there are distances and ranks, as well
  - gen (containing the spot numbers) as well as experim (containing the measurement IDs) are in the same format for easy handling)
- ngen x ncon (quality-) matrices:
  - pvalue - two-class pvalues
  - fpvalue - multi-class (i.e. only one per gene, but same size for compatibility)
  - minmaxseparation - min/max separation
  - stddevseparation - standarddeviation separation
- data annotation:
  - ginfo - gene annotations
  - anno - experiment annotations
Metadata

Variables in alphabetical order

*) Redundancy introduced for convenient handling.

name	size	class	content
backgroundsubtraction	1 x 13	char	'subtracted', 'not_subtracted', or 'not_available' (meta data)
ca	1 x 1	struct	recording last correspondence analysis (for updating figures)
casuppmode	1 x 1	double	determines which kind of supplementary points to plot in CA
ccthresh	1 x 1	double	correlation coefficient threshold for hybridizations (meta data)
closehybgaps	1 x 1	double	boolean: 0=plot hybs by measurement ID, 1=by download sequence (no gaps)
corranafig	1 x 4	double	handles of CA figures (2D, 3D, barplot, not used)
corranafigplotted	1 x 4	double	boolean register of which CA fig. is currently existing
displaysize	1 x 2	double	for maximal size of certain figures (e.g. overview fig.)
distprim	ngen x nexp	double	distance to control condition (analog to ratio), primary spotset
distsecu	ngen x nexp	double	distance to control condition (analog to ratio), secondary spots
experim	ngen*) x nexp	double	measurement IDs (as in database attribute "tableno") in ngen x nexp format (compatible with the large data matrices)
experiment_tables	1 x ncon	cell	measurement IDs by conditions
family	1 x ?	char	array family (type of chip), e.g. hb4
filterbytimepoints1	1 x ncon	double (logical)	boolean register of all conditions taken into account by filter 1
filterbytimepoints2	1 x ncon	double (logical)	boolean register of all conditions taken into account by filter 2
filterbytimepoints3	1 x ncon	double (logical)	boolean register of all conditions taken into account by filter 3
filtermode1	1 x 4	double	number representation of everything else selected for filter 1
filtermode2	1 x 4	double	number representation of everything else selected for filter 2
filtermode3	1 x 4	double	number representation of everything else selected for filter 3
fitconditionmedians	ngen x ncon	double	absolute intensity for each gene and condition (genewise medians of each condition within fitprim)
fitcondmediansratio	ngen x ncon	double	linear ratios for each gene and condition (fitconditionmedians/median(control measurements))
fitprim	ngen x nexp	double	normalized signal intensities (estimated for multi-channel data), primary spotset
fitsecu	ngen x nexp	double	normalized signal intensities (estimated for multi-channel data), secondary spotset
flatprim	ngen x nexp	double	(linear) ratios, primary spots
flatsecu	ngen x nexp	double	(linear) ratios, secondary spots
fpvalue	ngen x nexp*)	double	multi-class pvalue (f-statistic - based, computed by limma::eBayes; or imported from permutation test)
gen	ngen x nexp*)	double	gene ID number (database attribute "spotno"), large datamatix format
geneinfomode	1 x 1	double	way of representing genes in web browser (see view - settings - show what - geneinfo)
geneorder	1 x ngen	double	gene index (containing gene ID numbers out of {1..ngen} that survived filtering, sequence determined by sorting
ginfo	1 x ngen	cell	gene annotations
ginfoheader	1 x ?	char	attributes of gene annotations
histofig	1 x 4	double	handles of histogram figures
histofigplotted	1 x 4	double	boolean register of histogram figs.
histomode	3 x 5	double	number representation of View - Modes - Histogram Mode
linminmaxseparation	ngen x ncon	double	minmax sep. on linear abolute intensities (instead of log ratios)
linstddevseparation	ngen x ncon	double	stddev sep. on linear abolute intensities (instead of log ratios)
lowess_fraction	1 x 1	double	local window size for lowess normalization (default: 0.05)
medbool	ngen x ncon	double	1=upregulation, 0=downregulation
minmaxseparation	ngen x ncon	double	explained here in Fig. 5
multichannel	1 x nexp	double	hybridizations affiliated to the measurements, see here
myfig	0 x 0	double	contained fig. handle of browser figure, now empty
nexp	1 x 1	double	number of measurements
ngen	1 x 1	double	number of genes
normalization	1 x 48	char	normalization algorithm (meta data)
overviewfig	1 x 5	double	handles of overview figures
overviewfigplotted	1 x 5	double	boolean register of overview figs. existing at present
overviewmode	4 x 4	double	number representation of View - Modes - Overview Mode
partition	ngen x 2	double	first column contains gene IDs, second column the affiliated partitions: See here.
partitionmode	1 x 1	double	partition to be analyzed (0 = all)
port	1 x 4	char	port no. for inter-process communication
prim	ngen x nexp	double	raw data (primary spotset)
prim_bkg	ngen x nexp	double	local background as obtained from imaging software (primary spotset)
pvalue	ngen x nexp	double	two-class pvalue (computed by limma::eBayes; or imported from permutation-based test)
rankprim	ngen x nexp	double	ranks (primary spots)
ranksecu	ngen x nexp	double	ranks (secondary spots)
ratio	ngen x ncon	double	ratio (linear scale)
secu	ngen x nexp	double	raw data (secondary spotset)
secu_bkg	ngen x nexp	double	local background as obtained from imaging software (secondary spotset)
selection	1 x 1	cell	database structure 1 - style (zero padded) measurement IDs by MCE (multiconditional experiments in the browser selection). Normally, the data selected in the browser is merged into only one MCE before download. Therefore, this variable normally has size 1x1.
showwhat	1 x 4	double	boolean register of what to show (overview, zoom, quality, histogram). Select with View - Settings - Show What. Activate with View - Update View.
sortbytimepoints	1 x ncon	double (logical)	boolean register of all conditions taken into account for sorting
sortmode	1 x 2	double	number representation of everything else selected by Edit - Filters - Sort
stain	1 x ncon	double	number of measurements for each condition
stainb	1 x ncon	double	e.g. stainb(3):staine(3) yields column index of condition 3 in ngen x nexp - data matrices
staine	1 x ncon	double	see stainb (previous line)
statisticsfig	ncon x 5	double	handles of quality figures
statisticsfigplotted	ncon x 5	double	boolean register of quality figs. existing at present
statisticsmode	3 x 4	double	number representation of View - Modes - Signal Quality Mode
stddevseparation	ngen x ncon	double	explained here in Fig. 5
timepoint_is_control	1 x ncon	double	boolean register of control condition(s). Normally the first one (if not deleted, e.g. by Edit - Drop - Controls), only (for browser - selections comprising only one MCE)
x	? x 1	double	Selected genes (red circles and black text in CA fig.). Please note that x indexes an index: Gene IDs (= row numbers in a ngen x nexp - matrix) are obtained by geneorder(x) !!!
xblue	0 x 0	double	green gene tags, unlike x directly indexing gene IDs
xgo	0 x 0	double	selected gene ontology terms
xgreen	0 x 0	double	green gene tags, unlike x directly indexing gene IDs
xhyb	0 x 0	double	measurement tags (directly indexing measurement IDs (= columns in any ngen x nexp data matrix)
xred	0 x 0	double	red gene tags, unlike x directly indexing gene IDs
zoomfig	1 x 5	double	handles of zoom figures
zoomfigplotted	1 x 5	double	boolean register of overview figs. existing at present
zoommode	4 x ncon	double	number representation of View - Modes - Zoom Mode

Please note that depending on the kind of data loaded / functions invoked (e.g. using GO annotations), above list may be incomplete. If you stumble upon a variable that escaped documentation until now, please let me know.