This documentation is specifically for the knowledge objects that are used by CNS-TAP. For more information about the CNS-TAP application, see CNS-TAP Application.
Last Updated: April 28, 2022
KGrid Activator Version Used is: 1.3.6
Runtime Used: "V8" (JavaScript runtime called V8)
The Knowledge Objects in this repository collection are for the CNS-TAP Version 1.0 web application developed at the University of Michigan Medical School.
One original purpose of developing the CNS-TAP web application using Knowledge Objects was to have a manageable, extensible knowledge base to support the use of CNS-TAP in research and clinical trials in neuro-oncology.
This project requires two different Knoweldge Objects (KOs) that, when uploaded into CNS-TAP, enable computation of drug-specific and patient-specific scores for a variety of chemotherapy agents (i.e., drugs or medications). These scores indicate the likely usefulness of various chemotherapy drugs for patients with brain tumors.
All KOs have the following hierarchical structure of folders and files:
genericKnowledgeObject
|--metadata.json
|--service.yaml
|--deployment.yaml
|--src
|--payload_file_1
.
.
.
|-payload_file_n
Here is a picture of the file structure:
Every KO has: a metadata.json file with descriptive information including the version of the KO a service.yaml file describing an API or "service" supported by the KO a deployment.yaml file specifying how to run the code in the KO
Every KO also has one or more payload files. The payload files contain structured data or software code that machines can use. Generally, Payloads, whether in one or multiple files, contain biomedical knowledge that has been established through prior research.
There are only TWO KOs needed to run the CNS-TAP web application. They are described next.
The KO named "intrinsicDrugTable" is a special KO that outputs a table containing all of the “intrinsic” drug information for all drugs. This object is what holds all the intrinsic information about the drug that is not patient-specific. This includes the drug name, pathway, its score for in vitro data, its score for in vivo data, its score for safety, its score for CNS data, its score for brain penetration, and its score for FDA approval. Example shown below:
Dasatinib = {\
"**drug_name**":"Dasatinib", #drug name\
"**pathway**":"PDGFR", #drug pathway\
"tumor_in_vitro":4, #drug fact\
"tumor_in_vivo":6, #drug fact\
"phase_one_**safety_data**":6, #drug fact\
"cns_data_**with_**response":5, #drug fact\
"brain_penetration":10, #drug fact\
"fda_approv**al**":10, #drug fact\
"**total_points**":41, #total points\
},
Pazopanib_PDGFR = {
"drug_name":"Pazopanib_PDGFR",
"pathway":"PDGFR",
"tumor_in_vitro":4,
"tumor_in_vivo":6,
"phase_one_safety_data":6,
"cns_data_with_response":5,
"brain_penetration":5,
"fda_approval":10,
"total_points":36,
},
The KO named Tumor Patient Calculator is the second KO that contains the weighting for the three patient specific categories which are clonality, varient tier, and relevant clinical trial. These weightage values are what the app will multiply to the inputted values for these three columns in order to get the final CNS-TAP patient specific score. Example shown below:
var weights={
clonalityweight:5,
tierscoreweight:3,
trialweight:20,
}
From time to time, it is necessary to update the drug agent related content of the CNS-TAP application. This drug agent related content can be updated WITHOUT having to change the code of the CNS-TAP app.
To update this content, the following process can be followed.
The process will require downloading KOs from this GitHub repository, changing the content of those KOs, updating the version(s) of the KOs that are changed, packaging the KOs using special software, pushing the updated KOs back to GitHub, and then, finally, releasing the updated KOs in GitHub.
The process involves editing the content of one or both of the two KOs used by the CNS-TAP app. As described above, the two KOs are called, intrinsicDrugTable and tumorPatientCalculator.
Here are the steps to follow to download, edit, update version(s), package, upload, and release these KOs:
DOWNLOADING KOs from GitHub
-
From the command line of your local machine, clone repository into your local machine using this command:
git clone https://github.com/kgrid-objects/cnstap-objects.git -
To find one or both the Knowledge Objects needing to be changed, look in the collection folder.
EDITING KOs
- Update the contents of the Knowledge Object in any text editor of your choice, e.g., Atom (www.atom.io)
Next, three examples are given of making changes. The first two examples show changes made to the instrinsicDrugTable Knowledge Object. The final example shows changes made to the tumorPatientCalculator Knowledge Object. These changes are the anticipated types of changes that can be made without necessitating changes to the CNS-TAP app.
Here is an example of adding to the list of drug items in the "index.js" file within the SRC folder of the KO. Particular emphasis should be placed on the "editable" section because it is not directly a part of the table. The rule of thumb is that the "editable" score should be 1 if that given drug is the first drug in its respective pathway. Otherwise, the "editable" score should be 0. For example, if you are adding a new drug to the tool that is in a new pathway, this new entry will now be the FIRST and only drug in the new pathway so it will need an "editable" score of 1. However, if you are adding a new drug to a pre-existing pathway that already has several drugs in it, simply add the new drug to the bottom of the list of drugs within that pathway and give the new drug an "editable" score of 0.
Before
Sunitinib = {
"pathways": "PDGFR", "drugagents": "Sunitinib", "vitro": 2, "vivo": 6, "safety": 6, "cns": 5, "bbb": 0, "fda": 10, "subt": 29, "icln": 0, "itier": 0, "itrl":0, "cln": 0, "tier": 0, "trl": 0, "total": 29, "editable": 0,
},
Newly Added Content
,
Immune_Checkpoint_Inhibitors = {
"pathways": "PD-1", "drugagents": "Immune Checkpoint Inhibitors", "vitro": 4, "vivo": 6, "safety": 6, "cns": 10, "bbb": 0, "fda": 10, "subt": 36, "icln": 0, "itier": 0, "itrl":0, "cln": 0, "tier": 0, "trl": 0, "total": 36, "editable": 1,
}
After
Sunitinib = {
"pathways": "PDGFR", "drugagents": "Sunitinib", "vitro": 2, "vivo": 6, "safety": 6, "cns": 5, "bbb": 0, "fda": 10, "subt": 29, "icln": 0, "itier": 0, "itrl":0, "cln": 0, "tier": 0, "trl": 0, "total": 29, "editable": 0,
},
Immune_Checkpoint_Inhibitors = {
"pathways": "PD-1", "drugagents": "Immune Checkpoint Inhibitors", "vitro": 4, "vivo": 6, "safety": 6, "cns": 10, "bbb": 0, "fda": 10, "subt": 36, "icln": 0, "itier": 0, "itrl":0, "cln": 0, "tier": 0, "trl": 0, "total": 36, "editable": 1,
}
Before:
Sunitinib = {
"pathways": "PDGFR", "drugagents": "Sunitinib", "vitro": 3, "vivo": 6, "safety": 6, "cns": 5, "bbb": 0, "fda": 10, "subt": 30, "icln": 0, "itier": 0, "itrl":0, "cln": 0, "tier": 0, "trl": 0, "total": 30, "editable": 0,
},
After changing the entry for "vitro" from 3 to 2:
Sunitinib = {
"pathways": "PDGFR", "drugagents": "Sunitinib", "vitro": 2, "vivo": 6, "safety": 6, "cns": 5, "bbb": 0, "fda": 10, "subt": 29, "icln": 0, "itier": 0, "itrl":0, "cln": 0, "tier": 0, "trl": 0, "total": 29, "editable": 0,
},
Before:
var weights={
clonalityweight:8,
tierscoreweight:3,
trialweight:20,
}
After:
var weights={
clonalityweight:5,
tierscoreweight:3,
trialweight:20,
}
UPDATING KO VERSION(S)
After making changes to index.js files located in the src code folder for one or both of the knowledge objects, you will need to take the following steps to ensure that this changed index.js file code is correctly pushed and deployed into the CNS-TAP app running on the web. The first step ensures that the version of the index.js code is updated everytime a new change is made:
- Update version number within the @id, identifier, and version sections of the metadata.json file
Here is a picture of the metadata.json file which is where you can update version numbers:
Once you're done updating the versioning numbers, you're all set with regards to making changes to the actual KO content.
PACKAGING
Now, make sure to SAVE the edited index.js and metadata file or files and then open up the command line and follow the following steps:
-
At the command line (e.g., using Apple Mac Terminal) go to the "collection folder" containing the knowledge objects that you edited.
-
From inside the "collection folder", RUN
kgrid package-allNote: Kgrid CLI is needed to run this command. If it is not installed, then run this to install it:
npm i @kgrid/cli -g -
Next, RUN
kgrid create-manifestto create a new file manifest listing the ZIP files in the "collection folder."
UPLOAD CHANGED AND RE-PACKAGED KOs TO GITHUB
Now that you have run the commands that resulted in one or updated ZIP files with packaged KOs and an updated manifest, you will then need to make sure that the changes you made on your local machine are pushed to Github so that they can be deployed into the CNS-TAP app. Follow the next few steps to do this:
-
Push KO file changes to Github
-
Copy 3 files (manifest.json and the 2 zip files that have the name of each of the 2 objects) into the Github RELEASE section and upload the new files thereby replacing the existing 3 files with the new 3 files
RELEASE UPDATED KOs ON GITHUB
Here are pictures of the github release section on github and the name of the 3 files that need to be uploaded into the new release:
The github release section can be found on the right side of the repository below the section as depicted below
Then click on 'edit release' and replace the 3 files with the 3 new files as depicted below
- Finally, click 'update release' at the bottom and move on to the final stage of this task which is to trigger the KGrid Activator to upload the newly released Knowledge Object content.
TRIGGERING THE ACTIVATOR RUNNING IN HEROKU TO BRING IN THE NEWLY EDITED AND UPATED KNOWLEDGE OBJECTS
The final step involves downloading Insomnia, an API application, that will be used to ensure that the new changes are reflected in the app. Insomnia can be readily downloaded from their website: Insomnia Website
Once downloaded and installed, follow the following steps:
-
In this repository, go to the images section and download the file titled Insomnia-CNS-TAP.json Here is a link: Insomnia-CNS-TAP.json
-
Open the insomnia app, click the top left dropdown menu and then press 'Import/Export' as depicted below
-
Then, click the 'data' tab and upload the downloaded file (Insomnia-CNS-TAP.json)
-
Now, select the 'Reload' function on the left side and then click the 'Send' button as depicted below
- Then, select the 'Execute Intrinsic Drug Table' or the 'Execute Tumor Patient Calculator' button on the left side, depending on which knowledge object was updated. Click the 'Send' button as depicted below
At this point, the new changes should reflect in the code on the right. If not, you may need to redo the 'Reload' function a few additional times alongside the 'Execute' function untill the new changes appear in the code on the right.
Once the changes appear, the process is complete, and the Knowledge Objects are now fully updated. The CNS-TAP application should now reflect the updated objects.