Team:IISc-Bangalore/Hardware/Results
| Sponsors | |||||
|---|---|---|---|---|---|
 |
 |
 |
|||
 |
 |
||||
| Contact | |||||
|---|---|---|---|---|---|
|
|
|
|||
Calibration
Mini
Before we could try running growth curves with GCODe, we needed to check that its readings were linear. We calibrated the Mini (then called Version 1.1) against a spectrophotometer (Shimadzu UV-1800) at 648 nm by running serial dilutions of KMnO4 solutions and milk. We used a cuvette in the spectrophotometer, test tube in the Mini, and calbrated against cuvette reading with the Mini's electronic circuitry (the Cuvette Reading Graphs).
Data for these graphs available here.
As you can see, the graphs are linear to within experimental error!
Pro
The Pro is still under development, so we do not have much data from it yet. However, we did manage to measure its dilution capabilities:
As you can see, the Pro dilutes quite linearly, (once we correct for the offset) with an r-squared value of 0.996 !
(Data for this curve available at HW-Results-Pro-Dilution-Linearity-Data.xlsx)
Growth Curves
With the knowledge that the GCODe Mini's readings are linear, we are now ready to run growth curves! There is a slight hitch, however. Optical Density is calculated using Beer-Lambert's Law[1],
\[ \text{Absorbance} = \epsilon c l ,\\ \\ \text{ where}\\ \epsilon \text{ is Absorbance coefficient}\\ c \text{ is Concentration }\\ l \text{ is Path Length.} \]
But we realised that l and ε are different for our optical system as compared to the spectrophotometer. However since the concentration will still be the same in both cases, the OD will be proportional. We can then compare the graphs by normalizing, ie dividing by any one reading.
Growth curves of Salmonella typhimurium (STMwt14028) and Escherichia coli were made. The optical densities were simultaneously measured using a spectrophotometer at wavelength 600nm [In the graph as OD600] and our device GCODe Mini. [In the graph as V1OD_N_12H and V1OD_N_24H]. In V1OD_N_12H, the proportionality factor was found by dividing with the 12H reading, and in V2OD_N_24H with the 24H reading.
As you can see, the normalized GCODe Mini readings (red and cyan) match the spectrophotometer readings (green) almost perfectly!
Notice that the E.coli readings from the Mini form an even smoother curve than those from the spectrophotometer! This is probably because the GCODe readings were taken instantaneously, while it takes time to load the cuvette into the spectrophotometer and take a reading, allowing lots more opportunities for human error, leading to fluctuations in the actual time intervals.
At this point we were ready to try a growth curve with the GCODe all by its lonesome.
It worked perfectly! We only used a spectrophotometer at the very end, so that we could calibrate the curve to 'normal' OD units. (Data for this curve available at Ecoli-GrowthCurve-Solo1.xlsx)
To find out more about how our project evolved through the problems we faced, user testing and user feedback, head over to Hardware Evolution.
