For today’s group meeting, only the IRAM members (Xing, Robert, and Ashley) were in attendance. We primarily discussed the big picture scope of our project and the upcoming Photonics West presentations.
Experimental Talk
- Show that θmin of 10° vs 20° results in greater stability of predicted sizes over time. Analyze the data both ways and compare the fluctuations.
- Fixed cells vs live cells
- Single cells vs adding scattergrams of many cells
- Reduces the speckle noise and sparse sampling issues
- More like Foster data
- Zach averaged about 5 cells
- What exactly does this mean for time lapsed measurements?
- Reduces the speckle noise and sparse sampling issues
- Narrow vs broad distribution
- Not as big of a difference of a difference in fluctuation size between 10 and 20 for narrow distributions (like beads)
- Calibration/throughput vs angle
- For small scatterers or broad distributions, there are no “Mie features/ripples”, so the falloff slope is very important for extracting sizes
- Size very small beads (400-850nm). Can we fit them well?
- Take a “background” spectrum either with the laser off or with the laser on but no sample
- Lots of exposures to reduce noise
- Correct for data offset
- Also do this with the teflon cap
Simulation Talk
- Exploring the limits on the extracted size parameters’ reliability
- Measurement noise (from scattergram, including speckle)
- Model vs truth of parent cell (Gaussian vs log normal)
- Sparse distribution of sizes (sampling) vs smooth model
- How much does the predicted d50 fluctuate for different cells that actually have the same d50?
- Can we tell the difference between two cell populations by their d50?
- Shrinking the mitochondria over time- how will this affect the d50? Will all cells from the same parent cell have the same trend? Will their trends ever cross/overlap?
- Issue with sharp cutoff in the cross-section weighted distribution
- Either use a population with a smaller width, or make a larger lookup table