Example Alternative Growth Fit

Please first see the other Jupyter Notebook, explaining the basics of accessing mycelyso’s HDF5 files. Furthermore, this file assumes the output.h5 described in the other notebook to be present in the current directory.

Within this notebook, we will fit the mycelium length data using a third-party library, croissance (DOI: 10.5281/zenodo.229905 by Lars Schöning (2017)). Please install the current version off github first:

pip install https://github.com/biosustain/croissance/archive/master.zip

First, some general setup …

%matplotlib inline
%config InlineBackend.figure_formats=['svg']
import pandas
pandas.options.display.max_columns = None
import numpy as np
import warnings
import croissance
from croissance.figures import PDFWriter as CroissancePDFWriter
from matplotlib import pyplot

class OutputInstead:
    @classmethod
    def savefig(cls, fig):
        pyplot.gcf().set_size_inches(10, 12)
        pyplot.show()

# croissance's PDFWriter is supposed to write to a PDF
# but we want an inline figure, so we mock some bits
CroissancePDFWriter.doc = OutputInstead
CroissancePDFWriter._include_shifted_exponentials = False
def display_result(result, name="Mycelium Length"):
    return CroissancePDFWriter.write(CroissancePDFWriter, name, result)

warnings.simplefilter(action='ignore', category=FutureWarning)

pyplot.rcParams.update({
    'figure.figsize': (10, 6), 'svg.fonttype': 'none',
    'font.sans-serif': 'Arial', 'font.family': 'sans-serif',
    'image.cmap': 'gray_r', 'image.interpolation': 'none'
})

Opening the HDF5 file

We will load the output.h5 using pandas.HDFStore …

store = pandas.HDFStore('output.h5', 'r')
root = store.get_node('/')
for image_file in root.results:
    print(image_file)
    for position in image_file:
        print(position)
        break

/results/mycelyso_S_lividans_TK24_Complex_Medium_nd046_138_ome_tiff (Group) ''
/results/mycelyso_S_lividans_TK24_Complex_Medium_nd046_138_ome_tiff/pos_000000000_t_Collected (Group) ''

and load the first growth curve

result_table_collected = store[position.result_table_collected._v_pathname]
timepoint = result_table_collected.timepoint / (60*60)
length = result_table_collected.graph_edge_length

pyplot.title('Length over Time')

pyplot.xlabel('Time [h]')
pyplot.ylabel('Length [µm]')

pyplot.plot(timepoint, length)

[<matplotlib.lines.Line2D at 0x7fb7dd6e5160>]

Here, we will use the third party tool croissance to fit the data to an exponential growth model:

curve = pandas.Series(data=np.array(length), index=np.array(timepoint))

estimator = croissance.Estimator()
result = estimator.growth(curve)
# print(result)
print(result.growth_phases)

[GrowthPhase(start=9.928127173487246, end=22.594538504723342, slope=0.36693539043981077, intercept=3.1588539520729086, n0=-25.525547240977755, attributes={'SNR': 172.5009988033075, 'rank': 100.0})]

And furthermore use its plotting functionality to show the results:

print("Growth rate as determined by croissance µ=%.2f" % (result.growth_phases[0].slope,))
display_result(result)

Growth rate as determined by croissance µ=0.37