BCBP 400 Module 1: Problem Set

Name:___________________________________

Directions: Answer 5 of the following questions by inputting all answers into this word document so
that they can be easily read and parsed by the grader. Make sure to do problem #8 and turn in your own
work. Cite any work or phrasing that is not original. Do not share your electronic files with anyone. Goals
covered in this assignment: (A4-A8, M1-M2, C1-C2).

1. Reynold’s number. There are five situations described below. Calculate the Reynold’s number for each
situation by estimating the parameters in the equation. Is the situation a “low” Reynold’s number
situation where viscous forces dominate and we can ignore inertial forces?

a) Trapani’s zebrafish larva over in Biology

b) scallop

c) a chihuahua in the Great Molasses Flood in Boston

d) a flagella that rotates and moves without the cell body

e) same flagella as in c) that moves at the same speed, but it for some reason moves in a
direction that is perpendicular to the motion of the one in c)

2. Diffusion Coefficient. The Stokes-Einstein equation relates the diffusion coefficient D of a spherical
particle of radius a through a liquid with viscosity, :
D = kbT/ = kbT/6πa.
To better understand this equation let’s answer some questions about it.

a) What are the units for the diffusion coefficient? Take the Stokes-Einstein equation and
perform a unit analysis on it to prove your answer is correct.

b) What is ? At low Reynold’s number, if the particle were a rod-like flagella instead of a
spherical bead, would  change?

c) Estimate the diffusion coefficient for a small molecule like ATP, a globular protein like
calmodulin, and an organelle like mitochondria.
3. Swimming at low Reynold’s number. There are five animals described below. Can the animal swim at
low Reynold’s number? Why or why not?
a) animal with a rigid oar – Purcell describes an animal with a flexible oar, but what about an
animal with a rigid oar.

b) animal that has a screw on one side of it, which it turns in order to move

c) frog

d) jelly fish

e) mouse
https://www.youtube.com/watch?v=V5r1_qqDzcM

4. Propulsion Matrix. The propulsion matrix relates the force output by the propeller to the velocity
(rotational and translational) of the propeller. The propulsion matrix only depends on the propeller
shape and size and the viscosity of the fluid. That is, given a particular force, we would expect the exact
same velocity for a propeller with the same shape and size traveling the same direction through the
same fluid. Essentially, the propulsion matrix is telling you about the drag on the propeller—more drag
equals less velocity. So, we should be able to relate the propulsion matrix coefficients A, B, C, and D to
the Diffusion coefficients or the drag coefficients. See Djunta et al. for the derivation. One thing to note
is that the ABCD coefficients only depend on the translational diffusion coefficient along the direction of
movement (not perpendicular to it) and the rotational diffusion coefficient along the direction of
rotation (not the other two possible rotation axes). Let’s get a feel for the propulsion matrix. Let’s say
we have a flagellum moving through water with a propulsion matrix of ( CA DB ). Find the propulsion
matrix in the following instances. Use this other Purcell paper on propulsion as an aid.
https://www.pnas.org/doi/epdf/10.1073/pnas.94.21.11307

a) A flagellum moving through water with the same shape as the original, but twice as long.

b) A flagellum moving through water with the same shape and size as the original, but made of a
material that is twice the mass.

c) A flagellum moving through water with the same radius and length as the original, but twice
the amount of turns.

d) A flagellum moving through water with the same shape and length as the original, but twice
the radius.

e) A flagellum moving through water with the same shape, but twice the length and radius as
the original (2 times the size).

5. (A4, A6, C1, C2) There are two steps to particle tracking. The first step is to find the centroid of the
particle in a particular image. The second step is to locate the particle in multiple images and to link the
centroids together to form the track. In problem 6 you will use a program, MTrack2, that does all of this
quantification for you. So before you use this program let’s look “under the hood” and explore some
methods to find the centroid of a particle in an image. There is an image on our website under Module 1
called particle.jpg. Use this image to complete the following tasks.
a) One way to find the centroid is to use a fit. Open particle.jpg in IGOR or another data analysis
program and fit a 2D Gaussian to the image. In IGOR, make sure to “make” a matrix with the
right number of rows and columns and then set the matrix equal to the image values. You can
perform a curve fit on a matrix, but not an image. To fit the matrix, click on the checkbox “Show
Multivariate Functions” to get the Gauss2D function to show up in the “curve fitting” window.
Give the pixel location of the centroid in x and y and attach the image with the fit here.

b) The other way to find the centroid is to use a routine that does not require a fit. Here let’s
find the centroid by calculating the “center of mass”. First, use ImageJ to threshold particle.jpg
so that it is a black and white (binary) image. See how in problem 6. Now open the image in
IGOR and save the x and y locations of the white pixels in a wave. Calculate the average x and y
location of these pixels to find the centroid. List the centroid here and attach the binary image.

c) Which method do you think is the best in this case, and why? When would it be a good idea to
use the other method?

6. A4, A5, A6, A8, M1, M2) We model the intensity from a molecule as an Airy Disk.
a) What does an Airy Disk pattern look like? Include a graph here.

b) Why do we use an Airy disk as the model for the intensity for a fluorescent molecule? In your
answer make sure to estimate the size of a fluorescent particle and compare that to the
wavelength of light.

c) Derive Eq. 2 from Manzo et al.

7. (A4, A5, A6) We model diffusion inside of the cell as non-ergodic and anomalous.

a) What does anomalous diffusion mean?

b) What does non-ergodic diffusion mean?

c) Is diffusion in Fig. 2 of Djutanta et al. ergodic? Is it anomalous?

8. (A4, A6, C1, C2) To give you a better handle on how to measure the diffusion coefficient, D, using
particle tracking, let’s go ahead and analyze some Brownian motion data. I have placed some data on
our moodle website under Module 1. There is also a video in the ezLecture repository (ezLecture
Tracking) that walks you through how to open the data and track the particle using ImageJ. The data is
an image sequence of several 1.1-micron-diameter, fluorescent beads undergoing Brownian motion in
water at 23.6 degrees Celsius. The video is taken at 10 Hz (0.1 s per frame) and the images are at a
resolution of 160 nm per pixel. Follow the ezLecture and use ImageJ to output the file: trackresults. This
file will contain the x and y locations of the bead for each frame– this is the track. Your job is to take this
track and calculate the MSD as a function of time lag for one of the tracks. (This is the hard part.) Then,
you’ll plot this MSD vs. time lag. The following steps will walk you through the process. You can
complete them for one track or for all the tracks.

a) Plot a histogram of all of the x displacements over a time lag of one frame (0.1 s) and attach it
here. What is the average displacement? What is the average displacement supposed to be?
What is the variance? Use this variance to calculate the diffusion coefficient D. Note: in IGOR the
“width” of the Gaussian fit = sdev*sqrt(2).

b) Now we are going to calculate MSD’s. But, before we do, let’s do a practice. Here is a table to
calculate MSD by hand for the x and y pixel locations listed. Fill in the blank locations in the table
below or put the x and y data into Excel and import your table. You should get the MSD vs. time
lag graph below.
Frame x Delta x (lag = 1 frame) Delta x (lag = 2 frame) Delta x (lag = 3 frames) y Delta y (lag = 1 frame) Delta y (lag = 2 frames) Delta y (lag = 3 frames) SD (lag= 1 frame) SD (lag= 2 frame) SD (lag= 3 frame)
1 22 -1 0 1 345 -1 1
2 23 4 345 -1 -1 8 17
3 22 1 2 346 -1 0 -2 2
4 21 2 347 1 -1 5 2
5 19 -1 0 0 346 -2 -2 4
6 20 2 348 0 2 1 5
7 19 0 1 0 349 1 0 -1 1 1
8 19 1 0 348 -2 2 4
9 18 -1 349 -1 2
10 19 350

MSD (lag = 1 frame) MSD (lag = 2 frame) MSD (lag = 3 frame)

5
MSD (pixels^2)

0
0.05 0.1 0.15 0.2 0.25 0.3 0.35
Time lag (s)

c) Now calculate the two dimensional MSD for a time lag of one frame (0.1 s). Compare this to
the variance you found in a).
d) (OPTIONAL) Next, calculate the two dimensional MSD for multiple time lags from one frame
(0.1 s) all the way to 20 frames (2 s). Attach the graph here. The curve will not be a straight line
since there won’t be as much data at higher time lags.

e) (OPTIONAL) Fit a line to the MSD vs. time lag graph and use the slope to calculate the
diffusion coefficient, D, given the equation MSD=4 Dτ . Compare this to the actual diffusion
coefficient given by: D=kBT/(6*pi*eta*radius).