• Stars
    star
    117
  • Rank 301,828 (Top 6 %)
  • Language
    JavaScript
  • License
    MIT License
  • Created about 8 years ago
  • Updated almost 2 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

DCell browser and gene deletion simulator

DCell

A deep neural network simulating cell structure and function

Introduction

DCell is an application to provide an easy-to-use user interface and interpretable neural network structure for modeling cell structure and function.

Reference implementation is available here:

Publication

Using deep learning to model the hierarchical structure and function of a cell. Jianzhu Ma, Michael Ku Yu, Samson Fong, Keiichiro Ono, Eric Sage, Barry Demchak, Roded Sharan & Trey Ideker. Nature Methods, 2018

Directory Structure:

  • training/code: folder containing lua code for both neural network training and prediction.
  • training/TrainData: Training and predicting data.
  • training/Topology: Topology files for gene ontology.
  • backend: python wrapper code to perform predictions.
  • frontend: Javascript files to construct the web application server.
  • data-builder: Source files and scripts for backend database.

Dependencies:

The code is based on Lua Torch running on a GPU linux system. See here for installation and basic tutorials.

Demo

Training cmd:

th Train_DCell.lua -train training_file -test testing_file -topo ontology_file -save model_file

Predicting cmd:

th Predict_Dcell.lua -load model_file -test testing_file -out result_file [-gindex  gene_index_file]

Examples of training/testing files are in TrainData/ and ontology files are in Topology/.

Topology file defines the structure of an ontology as:

  • ROOT: term_name #genes
  • GENES: gene1, gene2, ...
  • TERMS: child_term1, child_term2

Output:

The model trained for each iteration will be saved in "-save model_file". The training program will produce a gene index mapping file saved in the same folder.

The predicting program will load both gene index file and trained model file and save the predictions in "-out result_file".

Data availability

To train the ontology on genetic interaction or growth using the gene ontology.

Please download the ontology at:

Genetic interaction and growth is at:

D-Cell predictions for Costanzo et al. 2010 dataset is at:

The running time on a standard Tesla K20 GPU takes <2 minutes for terms like "DNA repair", and 2-3 days for using the GO and ~7 millions training data.

User Documentation

Please visit our wiki.


Β© 2017-2018 UC, San Diego Trey Ideker Lab

Developed and Maintained by Keiichiro Ono (kono ucsd edu)

More Repositories

1

DrugCell

A visible neural network model for drug response prediction
Python
127
star
2

pyNBS

Python 2.7 implementation of network-based stratification (NBS) algorithm from Hofree et al (Nature Methods 2013)
Jupyter Notebook
37
star
3

MuSIC

Multi-Scale Integrated Cell
Python
35
star
4

TCRP

Few shot learning for cancer
Python
34
star
5

Network_Evaluation_Tools

Python 2.7 package with examples for evaluating a network's ability to group a given node set in network proximity.
Jupyter Notebook
32
star
6

cyREST

DEPRECATED. Please visit our new repository (cytoscape/cyREST)
Java
28
star
7

cy-rest-R

Example R script to use Cytoscape via RESTful API module.
HTML
27
star
8

hiview

HiView: the universal viewer for hierarchical data
JavaScript
16
star
9

heat-diffusion

Python
12
star
10

cy-rest-python

cyREST examples for Python users.
Python
11
star
11

llm_evaluation_for_gene_set_interpretation

Code space for 'Evaluation of large language models for discovery of gene set function'
Jupyter Notebook
9
star
12

cy-net-share

A simple web application to share network files generated with Cytoscape.
JavaScript
9
star
13

nexo

Prototype for NeXO web app.
JavaScript
8
star
14

tsri-lecture

Course material for TSRI network biology lecture
Jupyter Notebook
8
star
15

dot-app

Cytoscape application for exporting to .dot file format
Java
7
star
16

web.cytoscape

New version of cyNetShare
JavaScript
7
star
17

vizbi-2015

Sample data and notebooks for VIZBI 2015 tutorial session
5
star
18

jActiveModules

Java
4
star
19

cellmaps_pipeline

Python
4
star
20

cy-rest-node

Node.js examples for cyREST module.
JavaScript
3
star
21

cxmateold

A RESTFUL network API proxy service for network algorithms
Go
3
star
22

sdcsb-advanced-tutorial

Course material for SDCSB Advanced Cytoscape Workshop (4/17/2015)
3
star
23

auto-graph-visualizer

Automatic graph visualizer for the Cytoscape ecosystem
Python
3
star
24

neoelsa

Erlang
2
star
25

ce-components

Example implementation for CE component collection
TypeScript
2
star
26

large-graph-renderer

Webpack version of LGR
TypeScript
2
star
27

multitask_vnn

Multi-task learning VNN
Python
2
star
28

BiNGO

Java
2
star
29

ci-service-template

Template code to create new Cytoscape CI services.
Python
2
star
30

TreeViewer

D3-based DAG viewer for DCell web applicaiton
JavaScript
2
star
31

cyrest-examples

Latest example notebooks for CyREST
Jupyter Notebook
1
star
32

nest_vnn

VNN for drug response using NeST
Python
1
star
33

qfieldlayout

Python
1
star
34

obo-exporter

Jupyter Notebook
1
star
35

cellmaps_annotate_hierarchy

Jupyter Notebook
1
star
36

network-viewer

TypeScript
1
star
37

cdhidef

Python
1
star
38

webservice-ncbi-client

NCBI Client for CYtoscape 3. Moved from the core.
Java
1
star
39

ndex-web

React webapp code for CyNDEx2 Cytoscape App
JavaScript
1
star
40

diffusion-old

Heat diffusion daemon
Python
1
star
41

Automatic_graph_visualizer

Jupyter Notebook
1
star
42

cy-components

Monorepo for all React components maintained by Ideker Lab
JavaScript
1
star
43

drugcell-web-app

JavaScript
1
star
44

nexo-client

Client side module for NeXO web app.
JavaScript
1
star
45

cdoslom

Packaged OSLOM algorithm towards standardized community detection services for Cytoscape
Python
1
star
46

cellmaps_utils

Python
1
star
47

cxio_python

Python
1
star
48

ddot_rest_server

Python
1
star
49

NBGWAS-Frontend

A ReactJS Frontend page for the NBGWAS service created by Samson Fong and Dan Carlin
JavaScript
1
star
50

science-direct-app

Java
1
star
51

cyEZVis

Adding support for CalVR via Mugic plugin.
Java
1
star
52

ndex-valet-electron

NDEx Valet Electron app
JavaScript
1
star
53

GSAI

TypeScript
1
star