SR and DP Filtration

## [1] "Starting number of data points: 197439"
## [1] "Surviving edits: 22106"

Total Edits Permutations

Previously I had permuted by individual edits rather than by fish. I randomly assigned either Pop, Env or PopxEnv (groups) to each edit. As per discussion with Kim H, I think this is the wrong way to scramble. This treats each edit as an independent measure when actually the measure is total edits per fish. The variation I think that should be compared is the individual variation in total editing (total edits per fish), a null distribution would then be randomly assigning fishID (and carry over the Pop, Env and PopxEnv group designations), to the total number of edits from each fish.

E.g. Table of total sites per fish

Shows the first set of permutations, each column has 10 fish with a randomly assigned (replace = FALSE), number of sites

Set AHP AHNP NJP NJNP
1 589 589 471 507
1 373 373 664 741
1 737 737 964 445
1 710 710 879 539
1 599 599 527 326
1 419 419 263 502
1 737 737 337 316
1 546 546 424 729
1 843 843 632 206
1 329 329 706 558

For total sites, I altered the permutations such that each of the 40 individuals are randomly assigned to a total number of sites. I did not see any of the observed total sites from any group falling outside the null distribution. From this I was conclude there aren’t any group differences in total sites. However pulling out all common sites between the 4 groups might be interesting as the remaining sites will be unique to each group. My prediction here would be mismatched groups (AHNP and NJP), will have a larger set of unique sites…this would take awhile to code up the removal of common sites between groups. I think I will just present as total number of sites between groups does not differ. I also need to rerun total edits with the individual level permutations, the Pop and groups significance may then drop out.

Total Edits