mo105 / dreamteam Goto Github PK

-document how the schema was designed
-why they were designed that way
-what was used to create the database

-creating and maintaining a documentation of each of our part
-describing why we choose certain sites, methods etc

-documents what was chosen and why

I've drafted a new possible alternate phosphosite search page for everyone to see, just to make sure we're all on the same page. Please let me know if I misunderstood anything.

-when user upload a file, controller will create a list of dictionaries where substrate is the key, and phosphosite is the value
-create functions to return a list of kinase

For the user defined parameters, I have been thinking that it would be good to have this functionality:

• P-Value Threshold: where the user has the option to choose either 0.05 or 0.01 significance
• Coefficience of Variance Threshold (%): Either where the user can either input a value between 0 or 100
• Fold Change Significance Threshold: where the user can input a value between 0 and 5.

Do these sound okay?

Following my update on Sunday, the file app.py could not be ran on Windows. I've make a few amendments:

• Removed the auto generated init.py from app directory
• Removed the '.' from Database and forms in app.py
• Made a copy of Database/kinase_database.db in app (replacing the symlink because windows can't see it)

Then, I've tested on both Windows and Linux, using Powershell, Ubuntu (WSLv2), Ubuntu (Linux) and Pycharm. The following methods work fine when I tested them

• On powershell:
  python3 app.py

• On Ubuntu within Windows both of these worked:
  python3 app.py
  and
  flask run

• On linux both of these worked:
  python3 app.py
  and
  flask run

• On pycharm it works on Windows and Linux using the run button (provided the Working directory is set).

Sorry for the inconvenience

-documents the what's and why's for getting the kinase names, protein names and the aliases

def get_gene_protein_name(kinase_input):
"""
Returns a list of dictionary.
In the dictionary, there are gene name and protein name.
Returns empty list when no match is found.
>> kin = "AKT"
>> get_gene_protein_name(kin)
[{'Gene Name': 'AKT', 'Protein Name': 'RAC-alpha serine/threonine-protein kinase'},
{'Gene Name': 'AKT', 'Protein Name': 'RAC-beta serine/threonine-protein kinase'},
{'Gene Name': 'AKT', 'Protein Name': 'RAC-gamma serine/threonine-protein kinase'}]
"""
like_kin = "%{}%".format(kinase_input)
tmp = []
kinase_query = s.query(KinaseGeneMeta).join(KinaseGeneName).filter(or_(KinaseGeneName.gene_alias.like(like_kin), KinaseGeneMeta.uniprot_entry.like(like_kin),
KinaseGeneMeta.uniprot_number.like(like_kin), KinaseGeneMeta.protein_name.like(like_kin))).all()
for row in kinase_query:
results = {}
results["Gene_Name"] = row.to_dict()["gene_name"]
results["Protein_Name"] = row.to_dict()["protein_name"]
tmp.append(results)
return tmp
kin = "AKT"

-the part where the user upload a file and we need to display the relative activity visually

So I had a look at the sites that we talked about yesterday and at uniprot again. I have a few points which I'd like some input on.

1. Most sites do not disclose their databases openly.
2. Most sites do not have API that allows users to retrieve data from. There are ways to retrieve them but it's uber complicated.
3. Uniprot have API as well as all the information on point 1 (ie name, gene, where in cell, phophosites).
4. Even if we all use Uniprot, each of us could still work on getting something different then we could merge it using SQL afterwards (ie kinase & gene, kinase & location, kinase & family etc).
5. Uniprot has a lot of data. We should decide on what else to include.
6. The inhibitor sites allow download of the database, so it's not a problem.

Tldr: I think we should use uniprot for the kinase info part but we'll need to discuss which part we need. Inhibitor website is ok to export the data.

-document the how's, why's, what's and where's of getting the subcellular location for the kinase
-include a brief README.md to say which file is what

-when user search for something, the website would need to pull the information from the database
-more info on this: need to read around the subject

Question: In the search bar in the inhibitor page, are we allowing user to search an inhibitor or a gene? If we're allowing both, can we have a box that drop down next to it to specify inhibitor or gene?

-gene_name
-Uniprot_id
-Protein_name
-Gene_aliases

-documents the how's, what's, why's and where's of getting the information on inhibitors
-include a brief README.md in the folder to say what file is what

I'm going to pre-emptively create the functions for the following page.

Getting additional information for inhibitors so the final individual inhibitor page contains:
-inhibitor name
-inhibitor aliases
-empirical formula
-molecular weight
-smiles
-pubchem ID
-Inchi
-Image url
-target

I am in the process of making the project plan at the moment . So i was wondering does this look ok ?

Week 1:
Research (everyone)
Data mining (everyone)
Web Site Creation (Mo)

Week 2:
Data mining continue (everyone)
Database creation with SQL and organisation (Han)
Input software creation (Sheridan)

Week 3 :
Data mining continue (everyone)
Database creation with SQL and organisation Continue (Han)
Website creation (Mo)
- Anastasia : Help Page
- Task of Controller : Alisha
Documentation (everyone)

So I've drafted what we said about the phosphosite search pages. Let me know if this was what everyone had in mind.

-Finalize a list of human_kinases

def get_inhibitors_from_gene(kinase_gene):
"""
Take a string and return a list of dictionaries.
Returns empty list if there are no inhibitors.
>> get_inhibitors_from_gene("SGK1")
['GSK650394A', 'SGK-Sanofi-14i','SGK1-Sanofi-14g', 'SGK1-Sanofi-14h', 'SGK1-Sanofi-14n']
"""
results = []
kinase_query = s.query(KinaseGeneName).filter(KinaseGeneName.gene_alias==kinase).one()
for inhibitor in kinase_query.inhibitors:
results.append(inhibitor.inhibitor)
return results
kinase = "SGK1"
get_inhibitors_from_gene(kinase)

Final genomic location file has duplicated results and some data loss so I'm looking into fixing this.

Please go over sql

• Find where you can find this information
• Put this in a pandas dataframe

-Gene_name
-Uniprot_number(?)
-Cellular_location

-Gene_name
-Phosphosites
-Genomic_location
-Neighbouring_sequences

Hey, could everyone give me their respective column names? If you decide to change it at the end, it'll be fine, just send me a message.

Create the inhibitor csv.
Should contain the following columns:

• gene_name
• structure (as jpg or png in folder + filename in csv)
• molecular_weight
• empirical_formula

-create functions to query database for kinases

-creating the website

• [A]
• [B]
• [C]

-Gene_name
-Phosphorylation_sites

-documents the where's, when's, why's and how's of getting the details of phosphosites and substrates.
-include a brief README.md to say what file is what

-documents the how's, what's and why's of getting the genomic location
-include a README.md in the folder

creating the help page

create function to return all the information about an inhibitor.

Used when a user clicks on the inhibitor from the table. It will take them to the individual page for that inhibitor.

-using sqlite and sqlalchemy to create the database
-creating the functions to parse the info