notes

10-10-17
Initialized this repository after doing some work in it

What I have done is download the pairwise comparisons of fst between all populations into the fst folder. I have merged them into PBS with the two references and plotted that a couple of ways.

I also created a pairwise comparison file, converted all fst to zscores and used zkeep column to only keep windows at minimum represented by 20 reads (2000bp-ish).

I have also calculated pi, theta, tajimas D for neutrality in the angsd folder.
I plotted that several ways to look at global changes in pi, theta, and Tajimas.

I looked at admixture on all chromosomes and k2 and k4 were most supported so I plotted those.

added a keep section to pi_5kb

Read both zfst and zpi into R and subtracted zpi from zfst with a loop. Creates windows of interest for both high divergence and low diversity

10-17-17

Downloaded significant gwas that I ran over the weekend. Used bonferroni to correct individual alpha values to 4e-9. Took significant regions and converted them into .bed file.
merged individual SNPs into 5kb regions with bamtools merge.

10-19-17

Downloaded BioConductor and input each significant gwas SNP as an interval (1 indexed) formatr
things sort of worked in R, but I'm trying to do it with bedtools.
sorting .gff file with sort

grabbing the columns that are informative of the significant regions of divergence

11-02-17

Figured out the admixture/sources with fstadmix

working on using Noah's script to figure out zygocity of deletion in the /data/depth/...

11-29-17

Used awk to format neutral regions taken from 99% distribution of genomes:
cat list_neutral | awk '{OFS=""}{s=":"}{p="-"}{print $1,s,$2,p,$3}' > list_neut_final.txt

Used bedtoos multiinter (scripts/fst/overlap.sh) to find regions of overlap between all pops as well as unique regions in each pop.

12-7-17

Figured out a haploidizing function. Running that function with vcf created only from shared outlier regions between all populations.

zcat zshared.vcf.bgz | haploidizing_mine.r | ~/programs/htslib/bgzip > zshared_haplo.vcf.bgz

12-12-17
Using z regions - both shared by resistant and intermediate populations to look at distance between populations at those regions
converting them with awk to tabix format, so I can read them with xargs and isolate VCF with only those regions. From zinterm/res_out_all to z_formatted
'{OFS=""}{s=":"}{p="-"}{print $1,s,$2,p,$3}'

using zcat file | haploidize_gran_mine.r | bgzip > z_haplo.vcf.bgz

to haploidize the three groups of outliers

12-15-17

used bedtools to merge both pbs and z regions separately using the -o max,count option
then I used those windows and asked if the highest peak of the outlier falls as 1% outlier in each population. colored fig2 genomics from those results.

then I used 7 windows per peak in order to plot a mds plot without overrepresenting the huge region around AHR2a. plotted the shared regions.

12-18-17
I used a random set of regions the same size as the shared outliers for backgorund substructure (MDS).
grabbed them from the 5kb1kb.bed file and sampled at random with R sample()
converted to vcf, haploidized and plotted

Also plotted intermediate (selected by region and not 5kb) outliers and plotted those to update the graph.

Converted annotated outliers of shared regions with grep -f (file) (table) to get the Funhe nottation names

12-19-17

removed chr1 from analysis to see if mds distances are driven by it.

1-26-18

Did PBS on intermediates, comparing both with BB/GB and VB/SP. Nuthin! Good!

3-2-18

Added analysis of outliers vs each other. Interesting!
Both z values and PBS values.

4-26-18

re-subdividing wcfst into 1kb regions to have a direct comparison with the 1kb windows that noah calculated for Fst