Lab

Lab#

Learning Objectives#

At the end of this learning activity you will be able to:

  • Practice creating statistical figures to answer biological questions.

  • Practice writing figure legends for statistical figures.

  • Practice writing descriptive reasonings about a figure.

Note: It is difficult to automatically grade figures as they are many “correct” answers. So, most questions will accept any figure or axis and then ask you to answer a question that should be obvious from a properly generated figure. For all questions, assume a 95% interval.

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

%matplotlib inline

cell_level_data = pd.read_csv('pHrodo_DMEM.csv')
cell_level_data.head()

How full is each cell?#

The strategy of using the number of beads in a cell as our count is that it is impacted by the size of the cell. Small cells can only eat a few beads, large cells can eat many. To address this we’ll create a new measurement, the fraction of the cell containing beads. This way, small cells that are ‘stuffed’ with beads will beat out large cells with only a few beads.

For this analysis we’ll use:

  • ObjectAreaCh1 - The area of the entire cell.

  • SpotTotalAreaCh2 - The area of the cell containing beads.

Q1: Create an fraction_area_covered column#

Checked variables:

  • cell_level_data['fraction_area_covered'] - The fraction of the area of each cell covered by pH responsive beads

    • Calculate as red_bead_area / cell_area

  • q1_plot - A barplot showing the mean area of each cell covered split by well

    • Should be a matplotlib Axes object

Hint Create a new column by dividing the red_bead_area by cell_area for each cell. Then use groupby to calculate the mean fraction for each well and plot as a bar chart. See Module 7 walkthrough for examples.

Points

5

Public Checks

2

Hidden Tests

1

Points: 5

# What fraction of the cell's area is covered by phrodo beads

cell_level_data['fraction_area_covered'] = ...

# Create a barplot of the fraction_area_covered in each well
# Your answers should be between 0 and 1.

q1_plot = ...

grader.check("q1_area_covered")

Q2: Merge well_level_data with plate-map and visualize#

Checked variables:

  • plate_map - A DataFrame from reading the plate_map.csv file

  • well_level_data - A DataFrame where each well is an index (row) with columns: mean, sem, and count

    • Calculate from the fraction_area_covered of cells in each well using groupby

  • sample_level_data - A DataFrame merging plate_map and well_level_data

  • q2a - Which experimental condition (pHrodo_conc_ug) had less noise? (string or number)

  • q2b - Does this graph show evidence that dopamine increases bead percentage? (string: ‘yes’ or ‘no’)

  • q2_plot - A plot justifying your answers for q2a and q2b

    • Should be a matplotlib Axes object with error bars

Hint Use groupby on cell_level_data by 'well' to calculate mean, sem (using .sem()), and count. Merge with plate_map on the well column. Create a grouped barplot with error bars showing dopamine conditions and pHrodo concentrations. Less noise = smaller error bars. See Module 7 walkthrough for groupby aggregation and merging examples.

Points

5

Public Checks

2

Hidden Tests

1

Points: 5

# Load in plate map
plate_map = ...

# Group the cell level data by well and for each well calculate the mean, standard-error of the mean, and the number of cells

well_level_data = ...
well_level_data.head()

# Merge well_level_data with the platemap

sample_level_data = ...
sample_level_data.head()

### Visualize merged dataset
# Create any visualization which answers the questions below.
# Feel free to explore other functions like `lineplot` & `pointplot`.


q2_plot = ...

# Which experimental condition (pHrodo_conc_ug) had less noise in the measurement?
# Answer 5.0 or 7.5
q2a = ...

# Write your reasoning in a Markdown cell after this.

# Does this graph show evidence that dopamine increases 
# the percentage of the cell that contains beads?
# Anwser 'yes' or 'no'
q2b = ...

# Write your reasoning in a Markdown cell after this.

grader.check("q2_merge")

This week we explored how to summarize large datasets by sample. This aggregation is often important for downstream inferential tests like t-tests and ANOVAs. However, this technique also looses a significant amount of information; ~525 numbers are compressed to a single value. We will also explore more nuanced techniques like regression which allows us to use each of these points individually.

Submission#

Check:

  • That all tables and graphs are rendered properly.

  • Code completes without errors by using Restart & Run All.

  • All checks pass.

Then save the notebook and the File -> Download -> Download .ipynb. Upload this file to BBLearn.

Contents