These are chat archives for dereneaton/ipyrad

27th
Jan 2017
James Clugston
@Cycadales_twitter
Jan 27 2017 11:57
@dereneaton @isaacovercast Hi guys I have a problem with step 7. Thing seem to be working well generally however, when I change 21. min_samples_locus nothing is happen (I am rerunning it using the -f). I was using a population assessment file which I know will override that setting but I did remove the path to the file in the params. I am rerunning step 6 and 7 again now to see what happens.
Deren Eaton
@dereneaton
Jan 27 2017 16:06
Hey @Cycadales_twitter, is it that you are changing the min_samples setting in the populations file and nothing is changing?
James Clugston
@Cycadales_twitter
Jan 27 2017 16:10
@dereneaton I have now removed the path to population assignment file and tried just changing the min_samples and nothing is happening it seems stuck on the values set before using the opulation assignment file. I even tried changing the min samples to like 6 and still nothing. Also tried rerunning it from step six too with no luck. From my understanding this only affect step 7?
I hav attached my params file
Isaac Overcast
@isaacovercast
Jan 27 2017 16:11
Can you attach the json file as well?
James Clugston
@Cycadales_twitter
Jan 27 2017 16:12
yea sure
Deren Eaton
@dereneaton
Jan 27 2017 16:13
oh, @isaacovercast, I think it stores the self.populations dict when we parse the popfile, but we don't clear it if there is no popfile the next time, so it persists in the json object.
James Clugston
@Cycadales_twitter
Jan 27 2017 16:13
I am rerunning the analysis from step 4 now. To see if I can get it to change. I had not done this before
Isaac Overcast
@isaacovercast
Jan 27 2017 16:14
@dereneaton That was my suspicion.
Deren Eaton
@dereneaton
Jan 27 2017 16:16
@Cycadales_twitter looks like a bug. Should be easy to fix. Thanks for finding it. In the mean time, you could only apply the pop file to new branches in step7, and keep reusing the parent branch without the populations info to make new branches.
James Clugston
@Cycadales_twitter
Jan 27 2017 16:19
@dereneaton ahh ok I see. No worries. So I could just give the all the samples the sample population designation and then state what I want to keep?
Deren Eaton
@dereneaton
Jan 27 2017 16:23

I was thinking something like this:

## create an Assembly with pretty standard settings
ipyrad -n new

## enter data paths and stats into params and run steps 1-6
ipyrad -p params-new.txt -s 123456

## create a branch for an assembly that will use popfile 1
ipyrad -p param-new.txt -b pop1
## enter popfile1.txt into the popfile param and then run step7
ipyrad -p params-pop1.txt -s 7

## create a different branch for using popfile2
ipyrad -p param-new.txt -b pop2
## enter popfile2.txt into the popfile param and then run step7
ipyrad -p params-pop2.txt -s 7

That is assuming you want to try out applying different popfile settings to the same set of Samples. If you wanted to subsample the Samples then you could also remove some samples when branching.

James Clugston
@Cycadales_twitter
Jan 27 2017 16:33
For this dataset I would not think I need to subsample it. But I could give it a try but I did not want to start from scratch again.
But if I changed values in the population file would that also work?
Deren Eaton
@dereneaton
Jan 27 2017 16:34
It won't work right now because of the bug. But you can edit the JSON file by hand instead. Let me take a look and give you instructions.
OK, open the JSON file in a text editor and search for the line that says populations: {'name': val, 'name':val, ...} and remove everything inside the dictionary so it looks like this instead: populations: {},.
That will clear the populations file information so you do not have to start over to remove it. The next update will fix this so you do not have to do it yourself, and it will just clear when you remove the popfile path from your paramsfile. Does that make sense?
James Clugston
@Cycadales_twitter
Jan 27 2017 16:40
Yes I think that does make sense. I can see the defined population quite well. ''' "populations":{},'''
does this look right }, "name":"Ccal_85", "populations":{}, "database":"/mirror/1tb/james/Ccal/Ccal_85_outfiles/Ccal_85.hdf5", "_version":"0.5.13", "samples":[
Deren Eaton
@dereneaton
Jan 27 2017 16:40
yeah that looks right. Was it already empty?
James Clugston
@Cycadales_twitter
Jan 27 2017 16:41
no it was not
Deren Eaton
@dereneaton
Jan 27 2017 16:41
oh, ok. look good.
James Clugston
@Cycadales_twitter
Jan 27 2017 16:42
I just removed all the stuff. I might need to wait until this analysis stops running to try it but that helps loads. Any look with fineRADstructure
Deren Eaton
@dereneaton
Jan 27 2017 16:42
oh yeah, sorry not yet. It's on the list though.
James Clugston
@Cycadales_twitter
Jan 27 2017 16:47
No worries. But thanks for your help