Plot logTPM versus number of TFs bound (similar to presentation).
Plot lncRNA and mRNA seperate and Also include "reservoirs"

Here is the last little bits of 08 to finalize
08 GOOD SHAPE (need neighbor window analysis)
~55 there is something off with lncRNA starting out grey and then turning same pink :)
~93 label which plot is res non res -- add value etc
~223 label which is res or non res
~248 Anova or multi-group test for each promoter type
~284 Let's do the leave out res window and calculate neighbor window

Calculate a chi-squared of observed -vs - expected overlaps for mRNA and lncRNA promoters.
Calculate phi (effect size)
Plot Phi -vs - pvalue

I fixed most of the other counts in previous one and now just left with:

~60: quartile of 250 peaks filter.
~120: ks.test doesn't run
~255 The figure of peaks vs overlap for all promoters has intercept of 1,200 peaks?

****** 354 I did a length and summary and summary seems that min is 46 overlaps so no dbps don't bind any promters, but length is 161 ? ******* please double check

~364 need help with ggplot -- seems like there is a simply way to plot this with indexing lncRNA and mRNa overlaps in aes()

Plot the binding events in a density plot for all promoters
Subset into 4 density plots lncRNA, mRNA -vs -expressed, non-expressed
SImilar to slide 19 in presentation

make a density plot of number of binding events at a given promoter.
what DNA binding proteins did not overlap at any promoter?
what promoters never bind

Clean up to final .RMD and figure quality pdf

run the subset of DNA binding proteins from the presentation in Chipseeker (or collaborate with someone who has it installed)

Clean up to final .RMD

~146
#TODO this only saves last example eGFP-ZNF507

Perform Chi-squared test of binding at resevoirs versus non-reservoirs.
Calculate Phi

Plot P value versus Phi

Similar to slide 20

plot the number of peaks vs overlaps for lncRNAs, mRNAs in the same plot.
plot the slope of mRNA and lncRNA overlaps per number of total peaks

Clean up to final .RMD and figure quality pdf

09 : GOOD SHAPE
~56 the plot is a bit unwieldy just organize into columns of high, med and off
~84 Chi squared warning !
"Warning messages:
1: In chisq.test(df1) : Chi-squared approximation may be incorrect
2: In stats::chisq.test(x, y, ...) :
Chi-squared approximation may be incorrect
3: In chisq.test(df1) : Chi-squared approximation may be incorrect
4: In stats::chisq.test(x, y, ...) :
Chi-squared approximation may be incorrect

We have start values for promoters but no chromosome value :) probably need to add in 01 and repopulate across all directories :)

Clean up code of binary dendrogram as a figure quality pdf

Clean up to final .RMD

Remove the permutation test plots that are not used and ensure that all plots sync up with text.

Here are the minor remaining issues for 01 after careful curration :)

~# means that is about what line it's on -- they should be pretty accurate

~60: quartile of 250 peaks filter.
~120: ks.test doesn't run
~130 we could add peak widths from here?
~196 browser example of longest width window for RFX1
~255 The figure of peaks vs overlap for all promoters has intercept of 1,200 peaks?
~308 Print out min, max, median, mean, range:) of the #DBPs per promoter
LAST chunk doesn't run:
error:
333: Error: Failed to create output due to bad names.

• Choose another strategy with names_repair

Perform a genome wide permuation null distribution of binding events at each TE family and class.
Fisher Exact test of observed versus null.
Calculate P-value for each DNA binding factor
Calculate Z-score for each DNA binding factor
Make a clustered (family / class) heat map of Zscores

Similar chi-squared test now test

All ghosts versus
those with pol II and those with out pol II

Pol II Reservoir - vs- non Pol II reservoir
Bound -vs- unbound