1. Analysis was done on females with  samples
"GF_0169_F","GH_0081_M","GH_0087_F","FF_6136_F" removed,
so total sample size per group was:
5 FF, 6 GG, 6 HH
6 FG, 5 GF
6 FH, 6 HF
5 GH, 6 HG
(overall 51 mouse: 17 inbred, 34 F1s)

2. Files were taken from ftp server 
ftp://ftp.csbio.unc.edu/cegs_rnaseq_release3/
version of 5/8/2012.

3. Allele specific reads were produced using tag information for snps and indels.
For quality control the following settings were chosen:
average(base quality of the read)>=30
read map quality>=10
mask to keep specific reads = 0
mask to remove specific reads flag2rm=1796 
(based on samtools format specification http://samtools.sourceforge.net/SAM1.pdf
seep page 4)
read was considered to be coming from a particular allele if majority of the snps+indels (more than half) cam from this allele.

4. Top 10 genes by parent and strain are attached in the following files:
FG_parent.csv
FH_parent.csv
GH_parent.csv
FG_strain.csv
FH_strain.csv
GH_strain.csv
first two letters tell which cross was fit,
can be either parent of origin or strain effect.

for each file the column names are as follows:
gene_id - ensemble gene id 
short_name - short name of a gene
chr - which chromosome it it located on (all should be X)
start, end - gene positions
pval_strain - p-value of a strain effect
pval_parent - p-value of parent of origin effect
strain_eff - estimate of the strain effect 
(negative if expression of F is higher than expression of G,
if expression of F is higher than expression of H,
if expression of G is higher than expression of H )
parent_eff - estimate of the parent of origin effect (negative if maternal, positive if paternal)

columns:
GH_tot, HG_tot, GG_tot, HH_tot
GH_hapG, GH_hapH, HG_hapG, HG_hapH
- averages of normalised counts for each appropriate group,
first 4 averages for total read counts, then 4 averages for each allele specific read count.

5.  The model applied combined total read counts and allele specific counts as in a model as attached in a manuscript
http://www.bios.unc.edu/~zhabotyn/code/genes_2012_06_15/RNAseq.zip

6. This model includes estimation of:
strain effect (with test),
parent of origin effect (with test),
adjusted for individual total counts and Xce effect.