olaf.processing.graph_data_csv

Classes

GraphDataCSV

This class is called after the last image is reviewed in the GUI closes,

Module Contents

class olaf.processing.graph_data_csv.GraphDataCSV(folder_path, num_samples, vol_air_filt, wells_per_sample, filter_used, vol_susp, dict_samples_to_dilution, suffix='.csv', includes=('base',), excludes=('INPs_L',), date_col=False)

Bases: olaf.utils.data_handler.DataHandler

This class is called after the last image is reviewed in the GUI closes, and after the .csv file with the temperature ranges and frozen wells is created. It has a function that reads in in above-mentioned .csv file and calculates the INPs/L to use per temperature over all the dilutions for the experiment.

Parameters:

  • folder_path (pathlib.Path)

  • vol_air_filt (float)

  • wells_per_sample (int)

  • filter_used (float)

  • vol_susp (float)

  • dict_samples_to_dilution (dict)

  • suffix (str)

  • includes (tuple)

  • excludes (tuple)

vol_air_filt
wells_per_sample
filter_used
vol_susp
dict_to_samples_dilution
convert_INPs_L(header, save=True)

Convert from # frozen wells at temperature for certain dilution to INPs/L. The steps involved in this function are: 1. Seperate the temperature and # frozen well values. 2. Create a column with the total number of wells per temperature as affected by the background. 3. Calculate the INPs/L and the confidence intervals. It does this by using the number of wells frozen compared to the possible total number of wells. With the formula from <insert reference>: (INP/mL) = =(-LN((Dx-Ex)/Dx)/(Cx/1000))*Fx Dx = total number of wells minus the background Ex = number of frozen wells Cx = vol/well (microLiter) Fx = dilution factor 4. Prune the data by removing the INF’s and values that correspond with frozen wells This will allow for the logic in step 5 to function properly. 5. Combine the data into one dataframe, using logic that makes decisions comparing the last 4 values of a dilution before it has more than 29/32 wells frozen. The logic for this is:

<insert logic>

6. Save and return the data. The result is a dataframe with the temperature, dilution factor, INPs/L, and the lower and upper confidence intervals.

Parameters:

  • save – whether to save the data to a .csv file (default: True)

  • header (str)

Returns: the data as a pandas DataFrame

_error_calc(n_frozen, n_total, vol_well, dilution, z=Z)

Calculate the error of the INP/L The formula used is in (2) from: Agresti, A., & Coull, B. A. (1998). Approximate is better than “exact” for interval estimation of binomial proportions. The American Statistician, 52(2), 119–126. https://doi.org/10.2307/2685469 The formula is split up in three segments 1. The plus/minus part to differentiate between the upper and lower confidence interval 2. The remaining part of the numenator formula 3. The denominator of the remaining part

this calculates the values in INPs/mL and the typical conversion from mL to L applies for the errors too. :param n_frozen: number of frozen wells measured; single value or pandas df :param n_total: total number of wells; single value or pandas series :param vol_well: volume of each well; single value :param dilution: dilution (-fold); single value or pandas series :param z: z-value of the normal distribution, float

Returns:

error of the INP/L

Parameters:

  • vol_well (int | float)

  • z (float)

_INP_ml_to_L(ml_df)

Convert the INPs/mL to INPs/L, using the formula:: INPs/L = (INPs/mL * vol_susp) / (vol_air_filt * filter_used) with vol_susp = volume used for suspension vol_air_filt = volume of air filtered filter_used = proportion of filter used :param ml_df: dataframe containing the INPs/mL values. Could also be a single :param value or series:

Returns: same format as input, but with INPs/L values