Saving an epistasis model¶
All epistasis models/simulators store epistatic coefficients in Pandas Series/DataFrames, so the coefficients can be written to various file formats. This page lists a few.
Pickle¶
The recommended way to save an epistasis model to be used again is by pickling the model(See Python’s pickle
library).
# Import pickle module
import pickle
from epistasis.models import EpistasisLinearRegression
# Simple model object
model = EpistasisLinearRegression(model=1)
# Save to disk to open later.
with open('model.pickle', 'wb') as f:
pickle.dump(f, model)
To load the saved model,
# Import pickle module
import pickle
# Read from file.
with open('model.pickle', 'rb') as f:
model = pickle.load(f)
Warning
Pickled models will only work with the same version of the epistasis
package that created it. If you save a model and upgrade the library, you likely
won’t be able to use the model anymore.
Excel Spreadsheet¶
Epistatic coefficients can be written to an excel file using the to_excel
method
model.to_excel('data.xlsx')
sites | values | |
---|---|---|
0 | [0] | 0.501191 |
1 | [1] | -0.600019 |
2 | [2] | 0.064983 |
3 | [3] | 0.609166 |
4 | [4] | 0.242095 |
5 | [1, 2] | 0.286914 |
6 | [1, 3] | -0.264455 |
7 | [1, 4] | -0.464212 |
8 | [2, 3] | 0.638260 |
9 | [2, 4] | 0.235989 |
10 | [3, 4] | 0.717954 |
11 | [1, 2, 3] | -0.473122 |
12 | [1, 2, 4] | -0.041919 |
13 | [1, 3, 4] | -0.309124 |
14 | [2, 3, 4] | 0.606674 |
15 | [1, 2, 3, 4] | -0.268982 |
CSV File¶
Epistatic coefficients can be written to a csv file using the to_csv
method
model.epistasis.to_csv('epistasis.csv')
,sites,values
0,[0],0.5011910655025966
1,[1],-0.6000186681513706
2,[2],0.06498276930060931
3,[3],0.6091656756721153
4,[4],0.24209508436556937
5,"[1, 2]",0.2869142038187855
6,"[1, 3]",-0.26445514455225094
7,"[1, 4]",-0.4642116520437949
8,"[2, 3]",0.638260262428922
9,"[2, 4]",0.23598864236123118
10,"[3, 4]",0.7179538630349485
11,"[1, 2, 3]",-0.47312160287366267
12,"[1, 2, 4]",-0.04191888437610514
13,"[1, 3, 4]",-0.30912353449573415
14,"[2, 3, 4]",0.6066739725656609
15,"[1, 2, 3, 4]",-0.2689818206753505