README
-----------------------------------------------------------------------------------------------

* This directory consists of signal graph files, transcript abundance estimates versus genomic coordinate, for one of eight cell 
  lines generated by an analysis of the raw data produced from Phase 2 of Affymetrix/NCI Human Transcriptome project.
* A graph file consists 2 tab-separated values.

	* The first one being genomic coordinate.
	* The second one being its corresponding signal intensity value which corresponds to an estimate of 
	  transcript abundance at that coordinate.

* The genomic coordinates in these graphs are obtained after mapping the probes to Human Genome version 33 
  of NCBI. (April 2003 release)
* RNA samples were hybridized to 3 replicate arrays and analyzed together to create one signal graph file.
* Data for cell line 'SK-N-AS' is hosted here.
* Ten Chromosomes (6, 7, 13, 14, 19, 20, 21, 22, X, Y) span 98 chips.
* A graph is produced for every chromosome tiled on a chip 
  which can lead to multiple signal graph files for a given chip. For example, if chip 98 contains probes from chromosomes 19 
  and 20, then 2 graphs are generated for chip 98. 
* There are a total of 107 files in this folder.
* The filenaming convention for these files is as follows:
        * cellLine_chipnum_chr#.gz
        * SK_CHIP14_chr6.gz => SK cell line, chip #14, Probes on chip 14 are derived from chr6.

* The analysis methods used are as follows:

	* Replicate arrays were quantile-normalized and the median intensity (using both pm and mm intensities) of 
          each array was scaled to a target value of 44.
	* The expression level was estimated for each mapped probe position, i, by 

	      * collecting all the probe pairs that fell within a window of +/- 30bp
	      * calculating all non-redundant pairwise averages of pm - mm values of all probe pairs in the window
	      * taking the median of all resulting pairwise averages.

        * The resulting signal value is the Hodges-Lehmann estimator associated with the Wilcoxon signed rank statistic of the pm - mm 
          values that lie within +/- 30bp of the sliding window centered at every genomic coordinate