russodanielp / ciipro Goto Github PK

In addition to showing the CID identifier numbers in the CIIP Predictor results, also display the input-type as in the user uploaded file. More information shown in the results, like the chemical name, CAS number etc. might be helpful.

In the datasets module, the get_activities function loses data when the identifier is anything but cid

Create a pipe to import a dataset from pubchem.

Old user database was overwritten by the pull onto the ciipro server. lost old names. need to pull that database.

Currently, the MongoDB stores only a few PubChem Fields. More information (AID names, descrptions) can be added to the database.

Roll back changes when application breaks in middle of adding

Theres bottlebeck in the code calculating in vitro in vivo correlations of assays in profile to the in vivo activity (in the CIIProfile method in routes). Right now, it just iterates through each one and calculates the stats. This could easily be done by matrix multiplication and would likely be much faster.

make sure child compounds are getting deleted when removeing from the ORM

PubChem maps 'unique' identifers (i.e., CAS registry numbers) to several CIDs. Currently, only one is returned. Need to figure out a more appropriate rate to aggregate all that biological data.

check to make sure that user uploaded data does not contain duplicates, has activity values for all compounds, and that the identifier is the same for all and matches the selected identifier type (e.g, if someone uploads cids, make sure they are all cids)