griffithlab / deepsvr Goto Github PK
View Code? Open in Web Editor NEWLicense: MIT License
License: MIT License
Are there pre-trained models in the repository? After looking through the repo, I have found 3 instances of saved models:
Which one should be used if I do not plan to create and train the model from scratch. I could not find this information in the wiki.
Thank you!
Dear DeepSVR team,
Thank you for providing this tool for the community.
I have tried it on two cell lines with known mutation sites (majority are true mutations) and realized that the tool predicts every site to be a true mutation, and gives the exact same score (for ambiguous, fail, and somatic) for each of them. I first used your data to train a model, when that didn't work, I used other cell lines to train one, with similar results.
I also tweaked the number of epochs, and I got slightly different scores for different sites, but still everything was predicted as true mutations.
I'm wondering if you have an already trained model that I can directly use. I'm especially interested in your TCGA model (I have TCGA access), and whether that is available/could be made available under controlled access. If not, do you have any suggestions for training the model?
Thanks!
Wiki page https://github.com/griffithlab/DeepSVR/wiki/Create-the-Classifier contains links to nonexistent directory https://github.com/griffithlab/DeepSVR/tree/master/deepsvr/tests/test_data/create_classifier/
Hello,
First of all, thank you for the valuable work and detailed tutorial.
I was trying to install the Deepsvr package on Anaconda (OS: Window 64bit) and confronted with incompatibility issue that seems to be the problem within the package (e.g. I do not have packages such as 'matplotlib' or 'scikit-learn' installed in any of the environments, and packages such as 'convert_zero_one_based' seems to be original functions in Deepsvr). I tried its installation both directly through Anaconda Navigator Environment and by typing in the command in Anaconda Prompt (as written in the tutorial).
UnsatisfiableError: The following specifications were found to be incompatible with each other:
Package matplotlib conflicts for:
deepsvr -> matplotlib
Package h5py conflicts for:
deepsvr -> h5py
Package python conflicts for:
deepsvr -> python==3.6.1
Package seaborn conflicts for:
deepsvr -> seaborn
Package bam-readcount conflicts for:
deepsvr -> bam-readcount
Package click conflicts for:
deepsvr -> click
Package scikit-learn conflicts for:
deepsvr -> scikit-learn
Package pandas conflicts for:
deepsvr -> pandas==0.20.3
Package convert_zero_one_based conflicts for:
deepsvr -> convert_zero_one_based
Package numpy conflicts for:
deepsvr -> numpy==1.12.1
Package keras conflicts for:
deepsvr -> keras==2.0.4
Package tensorflow conflicts for:
deepsvr -> tensorflow[version='<=1.0.1']
In addition, I don't think Deepsvr runs on all Python3 versions; Deepsvr seems to be only supported on python 3.6.1. (I had this error message when installing so I had to make a separate environment for python 3.6.1 (my base is python 3.7)).
Hi,
I tried running DeepSVR by selecting around 300 raw mutect calls (i.e accepted and rejected), Of the 304 variants mutect ultimately accepted 45 and rejected 259, mostly because they were present in the normal. DeepSVR accepted (S) 291 of them and rejected (F) 15. I am using the default training set, is it expected that the default training should perform better than this or is it essential to develop a training set based on our own data? The ROC seems almost like it's inverting the labels, which is a little strange.
Also, a small side note, on the wiki page https://github.com/griffithlab/DeepSVR/wiki/Prepare-Data, it seems like the line 'If a header is present, it should list the "sample_name", "tumor_bam", "normal_bam", "reviewer", "solid_tumor", and "reference_fasta_file_path"' is missing the "Manual_review" header field.
Links to python notebooks on page https://github.com/griffithlab/DeepSVR/wiki/Data-Assembly give 404 file not found errors.
Broken links:
Is the sample.review.one_based file necessary for data preparation?
If so, how is it different from sample.review?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.