Scripts and results for simulated data
- casecon-noise.pl for Case/Control simulation
- quant-noise.pl for Quantitative outcome simulation
- sim_updated.R - R script for simulation, used for the results found here.
These are batch file scripts used to run on the compute cluster.
Simulate case-control data using sim_updated.R data and run it in PLATO (using each of the 5 different encodings) on the fly, resulting in data-all-1000
.
The nested structure of this folder is:
- MAF of 0.05, 0.1, 0.3, and 0.5
- Case-control ratio of 1, 2, and 3 combined with sample size of 2k, 10k, and 50k to get Case:Control simulations of:
- 500:1500
- 1000:1000
- 1500:500
- 2500:7500
- 5000:5000
- 5000:15000
- 7500:2500
- 10000:10000
- 15000:5000
- PENDIFF values of 0.25, 0.33, and 0.45
- 57 Simulation Models
- 6 main-effect models 6 encodings (REC, SUB, ADD, SUP, DOM, HET)
- 21 interaction-only models based on all combinations of those 6 models (named, for example like
ADD_x_DOM
) - 21 main effects models based on all combinations of those 6 models (named, for example, like
ADD_+_DOM
) - 8 models specified by penetrance tables (HR-HR, HR-HET, HR-HA, HET-HET, HET-HA, XOR, Hyp, RHyp)
- 1 NULL model
- PLATO results using 5 different encodings (add, dom, rec, codom, and weight (aka EDGE))
- Each result compressed in its own folder
- A text file reporting average pvalues across all simulations for each encoding
Simulate case-control data using sim_updated.R and run it in PLATO (using each of 5 different encodings) on the fly, resulting in data-SNR-1000
Results in data-SNR-1k-noMAF
are similar, but only at a MAF of 0.05.
The nested structure of this folder is:
- 5 different SNR values (0.01, 0.025, 0.05, 0.075, 0.10)
- 4 different MAF values(0.05, 0.10, 0.30, 0.50)
- 52 Simulation Models
- Same as above, except 18 main effects models and 19 interaction models, skipping:
- ADD_+_HET
- ADD_+_SUB
- DOM_+_DOM
- SUB_x_SUB
- SUP_x_HET
- Same as above, except 18 main effects models and 19 interaction models, skipping:
- PLATO results using 5 different encodings
Similar to RunSimAll.pbs, but using a single SNR of 0.05