Skip to content

DITTO using Neural Networks and SHAP#21

Merged
tkmamidi merged 144 commits intomainfrom
dev
Feb 2, 2024
Merged

DITTO using Neural Networks and SHAP#21
tkmamidi merged 144 commits intomainfrom
dev

Conversation

@tkmamidi
Copy link
Member

@tkmamidi tkmamidi commented Jan 6, 2024 •
edited
Loading

This PR includes enhancements to the repo for interpreting genetic variants from using traditional ML models to explainable Neural Networks. This branch includes the setup for local usage with installation details and the web application version of DITTO.

Here are the tasks you need to review:

Usage:

  • Review the readME:
    • Assess the resource requirements (CPU, storage, RAM) for typical use cases.
  • Webapp:
    • Confirm the webapp is accessible and functional at the provided link.
  • Local Setup:
    • Assess the resource requirements (CPU, storage, RAM) for typical use cases.
  • OpenCravat Setup:
    • Review the install_openCravat.md document for clarity and completeness.
  • Reproducibility:
    • Review the build_DITTO.md document for clarity and completeness.

This PR should close #8 #10 #20 .

tkmamidi added 30 commits April 26, 2023 16:33
@tkmamidi
first pass
b858e1b
@tkmamidi
until dbnsp
822e9f0
@tkmamidi
training config ready
dcbca0b
@tkmamidi
fixing clinvar config to get ID
b83a2c4
@tkmamidi
update col_config and python notebook for generating train and test f…
3039724 
…iles
@tkmamidi
training models
0abd693
@tkmamidi
fixing wierd instance with fatHMM values in parsing
658279e
@tkmamidi
changing NN script to use argparse instead of hardcoded file paths
937f1fa
@tkmamidi
adding DExTR to parsing stuff
442a3ab
@tkmamidi
adding DExTR file
9f34f68
@tkmamidi
parser can read and write
698dfe8 
.gz files
@tkmamidi
fixing duplicate columns in config
b7955a1
@tkmamidi
adjusting clinvar review column values
a2f9cd4
@tkmamidi
removing mitomap as there were no scores for training data
db66be8
@tkmamidi
Merge branch 'training' of https://github.com/uab-cgds-worthey/DITTO …
cb30221 
…into training
@tkmamidi
adding class weight to NN
f78fc1a
@tkmamidi
working DITTO NN
4dfbc59
@tkmamidi
concat training is working. need to test shap
794656b
@tkmamidi
DeepExplainer works but can't get shap_values. need to debug
7afa53b
@tkmamidi
DExTR scaled
b1cd211
@tkmamidi
NN on GPU and benchmarking for so
07d0435
@tkmamidi
added try:except to skip errors like sample size errors
a77063e
@tkmamidi
updated config file adding varity predictions
ad68913
@tkmamidi
adding more columns for clinvar parsing
7920c1b
@tkmamidi
new clinvar filtering with the actual clinvar column
74304ab
@tkmamidi
trying out GPU but it didn't work
2dadb5b
@tkmamidi
roc_auc and shap scores
6936244
@tkmamidi
split train and test by variants and retraining
1462faf
@tkmamidi
adds benchmark output as an excel sheet for all tools
fe3233f
@tkmamidi
fixing roc scores
453ed83
@tkmamidi tkmamidi self-assigned this Jan 6, 2024
ManavalanG
ManavalanG requested changes Jan 9, 2024
View reviewed changes
@tkmamidi tkmamidi removed the request for review from wilkb777 January 31, 2024 23:02
This was linked to issues Jan 31, 2024
package version differences for training and testing #10
Closed
Feedback on the pipeline #20
Closed
ManavalanG
ManavalanG reviewed Feb 1, 2024
View reviewed changes
ManavalanG
ManavalanG requested changes Feb 1, 2024
View reviewed changes
Copy link
Member

@ManavalanG ManavalanG left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good job, Tarun! Back at your court.

@tkmamidi tkmamidi merged commit 5c08367 into main Feb 2, 2024
@tkmamidi tkmamidi deleted the dev branch February 2, 2024 16:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@ManavalanG ManavalanG ManavalanG approved these changes

Assignees

@tkmamidi tkmamidi

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

Feedback on the pipeline package version differences for training and testing How does Ditto rank variants with same scores?

4 participants

@tkmamidi @ManavalanG @sdhutchins @wilkb777