array_fitting_tools

add to python path: array_fitting_tools/bin

Requirements (parantheses indicate what I am using):

Please update this list as you encounter dependencies not listed here!

• pandas (0.16.2)
• lmfit (0.8.3) ** Note that lmfit 0.9.0+ would NOT work with the current pipeline.
• joblib (0.9.0b2)
• seaborn (0.7.0)
• scikits.bootstrap (0.3.2)
• statsmodels (0.6.1)
• scipy (0.16.0)
• numpy (1.11.1)

scripts:

findSeqDistribution to annotate variants onto clusters.

To fit binding curves:

• processData to make series file from CPfluor directories.
• normalizeSeries to normalize fluoresecence by all-cluster images.
• singleClusterFits to do initial, minimally constrained fits on all single clusters.
• findFmaxDist to fit the distribution of fmax of good variants.
• bootStrapFits to revise estimates by constraining fmax and bootstrapping the fits.

To fit on/off rates:

• processData to make series file from CPfluor directories.
• normalizeSeries to normalize fluoresecence by all-cluster images.
• fitRatesPerCluster to fit individual clusters to on/off rates.
• bootStrapFitFile to bootstrap fit parameter to obtain 95% confidence intervals.

Other scripts:

• run_all_binding_curves.sh can be used to run all 4 scripts.
• compressBarcodes can take in a CPseq file and find unique barcode file. Also assesses quality.
• fitBackgroundTile will fit background clusters with the same method used above to better understand how noise contributes to fit values
  • This one probably won't work out of the box: talk to me.
• medianSubsetCPseries will take the per-variant median of a CPseries, and optionally subset to only include one tile.

More

For more info, see the wiki page