richelbilderbeek
Follow
richelbilderbeek
Follow
Pinned
5,201 contributions in the last year
Activity overview
Contributed to
richelbilderbeek/plinkr,
richelbilderbeek/nsphs_ml_qt,
richelbilderbeek/gcaer
and 198 other
repositories
Contribution activity
May 2022
Created 255 commits in 23 repositories
Created 3 repositories
Created a pull request in tresinformal/game that received 13 comments
key-action-map of player 1 can be saved and loaded, progress #522
This is some progress on #522: now the key-action-map of player 1 can be saved and loaded. For other players, this still fails, due to, for example…
+72
−5
•
13
comments
Opened 2 other pull requests in 1 repository
medical-genomics-group/gmrm
1
open
1
merged
Reviewed 4 pull requests in 1 repository
Created an issue in richelbilderbeek/nsphs_ml_qt that received 28 comments
[ANALYSE] Do a run on the many-SNPs data with for all combinations of autoencoder and phenotypic models
Use the data like #29 (IL-17RA), which as multiple SNPs
Todo:
Find best model combination: the M1 and p1 model combination gives the best results
…
•
28
comments
Opened 38 other issues in 6 repositories
richelbilderbeek/nsphs_ml_qt
17
open
8
closed
- Run with autoencoder-only for 2,3,5,10,25
- Do autoencode-only with latent layer of 3 neurons
- Plot in delta-AUC
- Plot in R2
- Plot end of phenotype predictions
- What is the relation between genotype concordance and R2?
- Do a long run with M1 and p1
- Determine if a SNP is kept in the autoencoder when used for prediction
- Compare with regular prediction: UMAP and linear regression
- Compare with regular prediction: PLINK PCA and fit linearly
- Compare with regular prediction: PLINK assoc_qt and betas
- [RUNNING] Filter out M and p models that do not work
-
Change the
trainingdir - Allow parameters to store more info
- Do a run on the many-SNPs data with only autoencoder functionality for all models
- Notes meeting Daniil 2022-05-12
- Completely separate internal/training and external/validation sets
- Do a SNP injection, to see if NN has learned genotypes
- Compare with regular prediction: latent layer and linear regression
- Compare with regular prediction: UMAP and linear regression
- Fix zipping
- Compare with regular prediction: PLINK PCA and fit linearly
- Always plot the distribution of the phenotype input data
- Remove GenoCAE folder from starter zip
- Remove labels