TODO

== BCD vs. MCD view ==

Create new fields for normalized data in the db and calculate and save them.

Make checkbox for switching between normalized and non-normalized

Put newest version on server (dev mode is OK for now).
  Run the three import and annotate scripts:
    import_google_docs_parts
    annotate_fpus
    annotate_plasmids

== Expression level ==

Normalized slider values.

Add "Download as GenBank" links.

Expanded promoter diagram.

== General ==

Script for removing the .0 on part names.
  Also modify the import_google_docs_parts to fix that bug.

Clean up excel sheet and make downloadable.

Add nice help page for each of the two views.

Add BIOFAB logo

Check that annotations are OK.

== Can wait ==

Add to SynBERC registry.

Put up:
  Attributions
  Link to article
  


  


-----------------------
OLD BELOW
----------------------

Friday:

- Talk to Vivek about:

    How should I calculate promoter and 5' UTR performance?
      This is for display in Expression Designer.

    Does the BCD vs. MCD view look ok?
      No: Should be GOI-mean-centered, not reporter-mean-centered
          Should be log2 before 0-100 normalization.
          Needs Y axis numbers and units.

    How should I calculate 5' UTR reliability?
      Just normalize the standard deviation to 0-10?
        meaning std_dev from BCD vs. MCD view?

- Import all plasmids as parts:
    script: import_google_doc_plasmids

- Annotate all plasmids:
    modify script: biofab-web-pilot/scripts/pilot_annotate
      Make it only run for plasmid parts.
      Make it delete existing annotations before running, for each part (plasmid).
      Make it only search for parts of type promoter, 5' UTR, GOI and terminator.
      Make it create real annotations instead of pilot-annotations.

    
Saturday:

- Calculate promoter and 5' UTR performance:
    Add a parts_values table and populate with values:
      performance
      performance_sd
      reliability (only for 
      reliability_sd

- Complete the Expression Selector!

    Fix data
      re-import data
      ensure no unwanted data is there
    Fix histogram (definitely wrong)
      add y-axis labels
    Better data in right-hand side.
    Show plasmid (copy code from bd_designs view).
    Make text-box for "must not contain".
    Remove "add constraint" button.
    Remove "add to cart" feature.
    Download as Genbank, plain text and SBOL.
    Download vivek's excel sheet link.

- Complete BCD vs. MCD view:
    add y-axis labels
    make it show actual sequence diagrams

Sunday:

  Complete the BCD-library view:
    BCD template
      copy from MCD vs. GOI view
    Sequence logo
    Straight bar chart.
    Click more info (not overlay):
      annotated plasmid
      level and sd in numbers
      BCD sequence diagram on mouse over bar chart
        copy from MCD vs. GOI view

  Add relevant links to top right side of pages

  Make front page links work.
  
Monday:

  Only if time:          
    Create PBD vs. BCD view
      3D bar chart
      Show EOU widget on click
        re-use from expression selector

  Add help information (do it in photoshop).


Tuesday:

  Put it in production mode.

  Make simple data-views for everything:
    Pilot [x]
    MPL and RPL [/]
    Terminator project [/]
    PBD [ ]
    BD-GOI [ ]
    Bicistronic library [ ]

If there is time:

  Terminator view:
    Bar-chart with one bar per terminator.
    Ask Guillaume what to show if one is clicked.

  RPL and MPL views:
    Copy BCD-library view.


------------------------------------------------------




Fix data:

  Questions for Vivek:

    1. Can I compare the BD-GOI measurements?
      Comparing a value between BD-GOI and PBD:
        BD-GOI: pTrc + BCD1 + GFP. Performance and std_dev: 365.98	34.77
        PBD: pTrc + BCD1 + GFP. Performance and std_dev: 462.95	68.40

    2. Are the FC averages in the spreadsheet log-scale?

    3. Should we use "advanced" performance measures for the 12 BCDs that have been tested more?
       Or should we just use the GFP performance from the "Randomized SD2 lib og BCD-GFP"-sheet for all?

    4. Why are there half-bicistronic designs in the PBD datasheet? (such as apFAB704)

    5. What's wrong with apFAB703 and apFAB704 in the google docs?

  Questions for Joao:

    1. Can I mean-center the standard deviations in the same way as the performances?

    2. Should I calculate separate scale-factors for GFP and RFP (between PBD and BD-GOI sets)
       or should I just make an average scale-factor.

  Fact: BD-GOI combinations all use pTrc promoter.

=== Maybe it's a good idea to do all of this in excel and just import after the fact! ===

  Designs imports both GFP and RFP values. Fix this:
    Import only all GFP. Skip performance values.

  Calculate normalized performance and standard deviation across GOIs (reliability):
    Calculate average performance of GOI across all BCDs.
    For each BCD-GOI combination, divide performance with the appropriate GOI-average.
      (if it is in log-space, subtract instead of divide)
      This will give the mean-centered performance for each BCD-GOI combination.
    Calculate the average of mean-centered performance for each BCD across all GOIs.
    Normalize the mean-centered performances to a scale of 0 to 100%.
      This will be the performance of a BCD (used when displaying the part diagram (OR NOT see Vivek question #3).
    Calculate the standard deviation of normalized mean-centered performance for each BCD across all GOIs (include GFP and RFP).
      This will be the reliability of a BCD.

  PBD calculations:
    Make the PBD values comparable to BD-GOI values:
      Two overlapping values (pTrc + BCD1 + GFP) and (pTrc + BCD1 + RFP).
      We want the BD-GOI equivalent for each of the two PBD values so:
        PBD_value * scale_factor = BD_GOI_value
        so: BD_GOI_value / PBD_value = scale_factor
      Calculate scale factor for GFP and for RFP and get average.
        (remember to use the non-mean-centered values when calculating scale factor)
      Multiply each PBD performance with scale_factor to get comparable values.

    Calculate the GOI-mean-centered performances of the Promoter-BCD combinations.
      (by dividing each of the performances with the appropriate GOI-average for the GOI used (either RFP or GFP))

    For each PBD calculate the average of the GOI-mean-centered performance of GFP and RFP (average of two).
      This is the per-PBD performance.

    Normalize the per-PBD performance to fit in a 0..100 range based on highest and lowest PBD performances.
        







Friday:

  Write script to cache annotations as JSON.
 
  Fix HTML+CSS for /designs

  Write script to import GOIs into new goi_designs table.

  Write script to calculate performance (standard error across GOIs) and save to designs table.

  Add a "must not contain" text-box constraint.
  Prototype of "extended info" on EOU widget click.

Weekend:

  Prototype the three views for the three bicistronic sub-projects.
  
Next week:

  Talk to Vivek to get 5' UTR annotations and make annotation-script.
  Talk to Guillaume about Terminator project view.
  Make one prototype view for bicistronic per two days.

Final week:

  Put up raw view for terminators and 5' UTR (1 day)

  Fix welcome screen (half day)
  
  Finish up expression picker (two days)

  Complete bicistronic design pages (1 days)

Last week before SynBERC:

  Get it running on the server (1 day)
  Testing and bugfixing (1 day)

----

Total todo:

  Create table for, and import, last sheet from Vivek's spreadsheet.

  Make three views for bicistronic sub-projects.

  Make simple "raw data" view (just copy pilot project view) for:
    5' UTRs (will likely need new table combining all three current tables)
    Terminators

  Fix welcome screen (look at drew's notes).

If time:

  Make nicer view for Terminator project.

  Make nicer view for pilot project.