spencersteers / ecmatlas Goto Github PK

• Protein
• Tissue
• ProteinHit
• Experiment
• VariableModification

Refer to the original database for some dummy data

HTML page to upload files

• Multiple file uploads
• Asynchronous uploads
• Read the files and leave a stub for the parser

Update proteins to use the cvs dataset we received. Add fixtures for tissue weight norms.

• Protein /proteins/
• Tissue /tissues/
• Experiment /tissues/
• ProteinHit /proteinhits/
• VariableModification /variablemodifications/

Example file of pulling wikipedia intro paragraph for a given protein.
http://cl.ly/code/3l1p0e442806

Will need to make a table for 'Editorial Content'. One to one relation with 'Proteins' that holds wikipedia information. Will need to store the summary and a link to the page on wikipedia.

Tissue weight norm
Dataset

Alongside the list of tissues a diagram of a mouse that the user can click on to select tissues.

Currently the Protein model is pretty plain. Most of the data comes from it's relationship to Experiments/ProteinHits.

Should have more general knowledge like:

https://www.proteomicsdb.org/proteomicsdb/#human/proteinDetails/52924/summary
http://en.wikipedia.org/wiki/MYH6 (sidebar)

Currently the Protein model doesn't have clear relations to organs. Have to go through experiments which is silly.

Programmatic access guide: http://www.uniprot.org/help/programmatic_access

Example query for gene col1a1 to get id: http://www.uniprot.org/uniprot/?query=gene:col1a1+AND+organism:rat&format=tab

Look up how to make http queries in python and how to iterate through a tab delim string.

• Link proteins in database to a uniprot identifier
• Provide a link to Uniprot for each protein
• Fetch Uniprot protein descriptions and insert into database (might be done with wikipedia now #16 )
• Allow for fetched Uniprot protein descriptions to be changed and updated (might be done with wikipedia now #16)

Set up the Django admin interface so that the tables are available for editing. Start with small models like Tissue, Family and FunctionalGroup. Eventually the admin will need to be updated to allow editing of the Uniprot (#15) and Wikipedia (#16) data.

New fields and tables will be needed. Experiment table with a date and name. Dataset table will need a foreign key relationship to Experiment.

All data loading happens in ecmdatabase.views.py

CSV parsing is handled in ecmdatabase.dataset_parser.py. This is where column names would be changed to match the new dataset.

The template being rendered is ecmdatabase/templates/upload.html and the loader can be viewed at /datasets/upload/. I'd start by displaying the upload date of a dataset on this page.

• Add support for experiment ID, experiment date, and relative concentration fields to dataset parser
• Allow for multiple datasets to be grouped under an experiment ID
• Add experiment date to data loader
• Display dataset upload date when viewing list of uploads in the data loader