Column definition for segments.txt files
----------------------------------------

chr: Arabidopsis chromosome on which TAR is located

start: Start position of the TAR (first nucleotide of the most 5'
probe inclusive to the TAR)

stop: End position of the TAR (last nucleotide of the most 3' probe
inclusive to the TAR)

num_probes: Number of tiling probes inclusive to the TAR

level: Predicted discrete expression level of this TAR (between 1 and
10)

repeats: Fraction of repetitive probes inclusive to this TAR (repeat
annotation is based on degenerate 25-mer repeats in the A. thaliana
genome, see Clark et al, Science, 2007 for details)

high_confidence: TARs with at least 4 probes, expression level 6-10
and repetitive proportion less than 0.25 are called 'high-confidence'
TARs.

novel: 0 if there is an overlap of at least 25 nt to an EST or cDNA,
else 1 (ESTs and cDNAs were downloaded from the TAIR ftp server,
release 7)

genic: 1 if this TAR overlaps (by at least 1 nt) to a gene annotated
in TAIR7, 0 otherwise


Column definition for unannotated_segments.txt files
----------------------------------------------------
same as for segments.txt files

content is a subset of segments.txt extracted according
to the following criteria:
i) TARs have to be unannotated
ii) TARs have to be high-confidence predictions

TARs are sorted in descending order of the number of probes and the
predicted expression level.



Column definition for stress_segments_stat_test.txt files
---------------------------------------------------------

chr: same as in segments.txt files

start: same as in segments.txt files

stop: same as in segments.txt files

num_probes: same as in segments.txt files

level: same as in segments.txt files

upstr_gene: AGI code of the nearest annotated gene upstream of this
TAR

downstr_gene: AGI code of the nearest annotated gene downstream of
this TAR

upstr_gene_dist: distance in nucleotides to the nearest annotated gene
upstream

downstr_gene_dist: distance in nucleotides to the nearest annotated
gene downstream

p_value: p-value of a Mann-Whitney U test for significantly increased
signal in treatment vs. control arrays (as indicated by the file
name). The test compares the median intensities (pooling replicate
arrays) of all probes inclusive to this TAR between treatment and
control. Lists are cut off at the 5% alpha level.

repeats: same as in segments.txt files

conservation: sequence conservation of this TAR compared between
Arabidopsis, rice, poplar and Sorghum assessed as described in Methods

high_confidence: same as in segments.txt files

unannotated: same as in segments.txt files

novel: same as in segments.txt files

genic: same as in segments.txt files

overlapping to segments of other samples (S_001-S_013): thirteen
columns, one for each sample (mock controls as well as treatments)
with value 0 or 1. If a TAR overlaps to TARs predicted for other
samples this is indicated by a 1 (0 else).