Experiment 2: Discussion
Sources of Error
Sources of error in the second experiment include the determination of the total number of transformants in a plate based on the number of CFU in several samples from that plate. These samples can not exactly represent the entire plate. A ranking of plates based on a visual estimation of colony density was also done, to be compared with the average number of CFU’s counted per centimeter square. The calculated values did agree with the visual estimations in all but three cases: pH 5.00-6.01/A, pH 6.88-7.14/A and pH 8.98-9.84/A, for which the calculated value was too low. The biggest discrepancy, for pH 8.98-9.84/A, was about 25% different from the visually estimated value. Thus, these errors are major, but do not make the numerical data extraneous. A more diluted cell suspension could have been plated on ampicillin, such that all CFU’s could be counted. However, the dilution process would also introduce major error. Other, unavoidable counting errors had only a minor impact. Colonies varied in size, and the limit of when they were considered too small to count slightly varied slightly while counting. Some colonies had also merged into each other and may have been counted as more or fewer CFU than actually present.
Sources of error in competence development from the first experiment still apply to the second experiment, as the same competent cells were used in both experiments.
Serial Dilution Error
Serial dilution error is expected to be much lower than in the first experiment, as far more similar numbers of CFU’s on plates of the diluted suspension were observed in this experiment than the first experiment. Each suspension was mixed more thoroughly than in the first experiment. The lower calculated number of cells plated for the sample at pH 8.62-9.30 can be accounted for by a human error before serial dilution. The variation in the number of bacteria calculated to have been plated may be a result of different survival rates of bacteria at different pH levels, or, most likely, random errors throughout competence development and transformation.
Inverse Correlation
The transformation frequency still varies somewhat inversely with the number of bacteria plated, although less strongly than in the first experiment and only at pH levels lower than 7 or greater than 8. Since transformants/cm2 was calculated without using the number of bacteria plated, the similarity between transfromants/cm2 and transformation frequency graphs suggests that the remaining inverse correlation between transformation frequency and the number of bacteria plated is not solely a result of remaining serial dilution error. This evidence supports Norgard’s findings of an inverse correlation between the number of bacteria plated and the transformation frequency. The more bacteria are plated, the lower a transformant’s chance of survival and the lower the transformation frequency becomes. This suggests error from variation in the number of bacteria plated remains present in calculations of transformation frequency. Population density is more closely accounted for in the third experiment analysis.
Optimal pH
There is a 97% correlation between the sections of the first and second experiment graphs from pH 5.5 to 8. In the second experiment graph, a peak at pH 7 is evident in both transfromants/cm2 and transformation frequency graphs, suggesting these observations are not solely a result of a serial dilution error. The decrease in frequency by pH 8 cannot even be explained as a result of an inverse correlation between the number of bacteria plated and the transformation frequency. The drop in transformation frequency by pH 8 is therefore is therefore in response to extracellular pH. As in the first experiment, the transformation frequency peak at pH 6.88-7.14 is similar to Norgard’s findings that transformation frequency is greatest at pH 7.25-7.75, but occurs at a slightly more acidic range. Transformation frequency is tested at smaller increments between pH 6 and 8 in the third experiment, to determine a more specific optimum pH.
PMF involvement
The drop in frequency by pH 8 supports that neither component of the PMF is involved in E. coli transformation, as discussed with the first experiment.
pH 9.3
According to the graph of second experiment results, transformation frequency is greatest at pH 9.3. This is largely a result of the much lower number of bacteria plated on these plates, due to the inverse correlation between the number of bacteria plated and transformation frequency. There is, however, the unlikely possibility that the peak at 9.30 is not solely a result of error. Thus, bacteria were made competent and transformed again at approximately pH 9, in the third experiment.
