Skip to content

ML #181

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
jessicaw9910 wants to merge 281 commits into
base: main
Choose a base branch
from
Open

ML #181

jessicaw9910 wants to merge 281 commits into from

Conversation

@jessicaw9910
Copy link
Collaborator

@jessicaw9910 jessicaw9910 commented Nov 11, 2025 •
edited
Loading

Description

Sub-package for ML; currently sequence focused

Todos

Notable points that this PR has either accomplished or will accomplish.

  • Process raw PKIS2, Davis data
  • Load PKIS2 data
  • Load Davis data
  • Drug models
  • Kinase models
  • MLP on pooling layer
  • Cross-attention on full latent embedding (IN PROGRESS)
  • Factor model for mutational data (IN PROGRESS)
  • Factor model
  • Tokenizers
  • Configs
  • Update Davis instead of using TDC
  • Standardize SMILES strings
  • Use _deserialization_cache method for DICT_KINASE
  • Use PyTorch CUDA expandable segments
  • Absolute not relative paths in batch scripts
  • Cluster kinases and drugs for splits
  • Docs
  • Tests

Questions

  • Want to perform a per vector dot product rather than a matrix multiplication - how? Einsum
  • How to batch if using CV dataset from Python? Set up utils_trainer.py script that batches cross-validation jobs

Status

  • Ready to go

jessicaw9910 and others added 30 commits April 17, 2025 15:22
@jessicaw9910
Merge branch 'main' into ml
e83845e
@pre-commit-ci
[pre-commit.ci] auto fixes from pre-commit.com hooks
22f5c2e 
for more information, see https://pre-commit.ci
@jessicaw9910
switched from pkg_resources to importlib since fully deprecated in 3.12
91e3f85
@jessicaw9910
4/18 update
06d6ec8
@jessicaw9910
Update requirements.yaml (#114)
5d17492 
Try pinning Python version (python >=3.10,<3.11)
@jessicaw9910
Update conf.py (#115)
5408564 
Fixed import statements
@jessicaw9910
Update index.rst (#116)
4ea4bb2 
Increased max depth from 3 to 5
@jessicaw9910
Update conf.py (#117)
e172301 
changing sys.path
@jessicaw9910
Update api.rst (#118)
4ccd417 
Using sub-modules in API
@jessicaw9910
Update api.rst (#119)
6cf5357 
Removing api from submodules
@jessicaw9910
Update api.rst (#120)
3e8c36f 
Removing sub-modules
@jessicaw9910
Update index.rst (#121)
feee710 
Changed max depth to 6
@jessicaw9910
reproducible conversion of PKIS accession to UniProt ID
45b80e3
@jessicaw9910
Merge branch 'main' into ml
9719c4d
@pre-commit-ci
[pre-commit.ci] auto fixes from pre-commit.com hooks
42cc1c0 
for more information, see https://pre-commit.ci
@jessicaw9910
script to generate harmonized PKIS2 data
34c777f
@jessicaw9910
added get_repo_root to utils
a6f767a
@jessicaw9910
updated pkis2 annotated file for latest reconciliation
5df6bf7
@jessicaw9910
created a FineTuneDataset class to load data from csv files, split on…
855fc59 
… kinase group, transform data with StandardScaler, and tokenize; also created a more specific PKIS2Dataset class
@jessicaw9910
training scripts and initial pooling model
61431d0
@jessicaw9910
added run_trainer as a CLI
105905b
@jessicaw9910
log_config for mkt.ml
0422b73
@jessicaw9910
clustering module
f2d88ac
@pre-commit-ci
[pre-commit.ci] auto fixes from pre-commit.com hooks
284b532 
for more information, see https://pre-commit.ci
@jessicaw9910
Merge branch 'main' into ml
ce79c60
@jessicaw9910
keep only latest
459da92
@jessicaw9910
uncomment run_pipeline_with_wandb for pre-commit CI
054edff
@jessicaw9910
commenting out all of models file - no longer in use but preserve for…
d20fe2d 
… posterity
@jessicaw9910
pre-commit ci
e68ba82
@pre-commit-ci
[pre-commit.ci] auto fixes from pre-commit.com hooks
78cba29 
for more information, see https://pre-commit.ci
jessicaw9910 and others added 29 commits October 8, 2025 11:17
@jessicaw9910
Merge branch 'ml' of https://github.com/choderalab/missense-kinase-to…
21fea35 
…olkit into ml
@pre-commit-ci
[pre-commit.ci] auto fixes from pre-commit.com hooks
1a23c53 
for more information, see https://pre-commit.ci
@jessicaw9910
merge main
bf994a3
@jessicaw9910
Merge branch 'main' into ml
84f7c37
@jessicaw9910
merge main
91fe431
@jessicaw9910
Merge branch 'ml' of https://github.com/choderalab/missense-kinase-to…
a25d63a 
…olkit into ml
@jessicaw9910
deleted unused missense_kinase_toolkit/ml/mkt/ml/datasets modules tha…
629d0a3 
…t have been moved to mkt.databases
@jessicaw9910
updated constants for new variable kinase input columns
fb86a1d
@jessicaw9910
clarified kinase/drug input (i.e., not names); add bool_wt - need to …
4c8ceb6 
…implement
@jessicaw9910
updated PKIS dict for new arguments
cb8d0fa
@jessicaw9910
add config to output file
77a69fb
@jessicaw9910
removed PKIS2 since no longer necessary - everything pre-processed in…
37bc9f5 
… mkt.databases.datasets; fintune now contains KinaseSplit and CrossValidation classes
@jessicaw9910
harmonized with finetune changes; KinaseGroupSource removed
009e1ca
@jessicaw9910
added KinaseInputType
db63519
@jessicaw9910
removed dataset sub-folder and moved finetune to outer folder
0565ebc
@jessicaw9910
updated finetune module changes
2d5e72a
@jessicaw9910
KinaseInputType to StrEnum in constants
3a7040f
@jessicaw9910
need to drop None self.df[self.col_drug] values before tokenization o…
dba57ac 
…r throws an error
@jessicaw9910
added info flag for dropna
e194e96
@jessicaw9910
save metrics to csv and plt.imshow heatmap option for predictions rat…
09c79e2 
…her than just a scatter
@jessicaw9910
switched from componc_gpu to componc_gpu_batch partition as default
bc8c78b
@jessicaw9910
fixed partition in CLI
81f634d
@jessicaw9910
add config to sub-directory for CV (previously only did it for train-…
e3d7efc 
…test)
@jessicaw9910
fixed pred vs. actual labels on scatterplots/heatmaps
bc0dfc9
@jessicaw9910
updated plotting functions for ml sub-package
7567ed9
@jessicaw9910
updated dynamic range plot - update for new data when ready
a1f0722
@jessicaw9910
added PYTORCH_CUDA expandable segments flag to batch script; absolute…
b23e54e 
… paths in bash script and check script_dir exists so can resume run anywhere; lowercase comments
@pre-commit-ci
[pre-commit.ci] auto fixes from pre-commit.com hooks
0c85140 
for more information, see https://pre-commit.ci
@jessicaw9910
Merge branch 'main' into ml
bf7f2cf
@jessicaw9910 jessicaw9910 changed the title Ml ML Nov 12, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@jessicaw9910