Computational Analysis


We developed a computer program to convert the tumbling of E. coli into audible frequencies. We visually analyzed our videos of varying conditions and determined the rotations per second of 5 randomly selected tumbling bacteria. An array with this data is entered into the program along with the average length of the cells being classified (usually ~2 um). The program uses this information in a model that assumes a 2-D plane of tumbling to calculate the flagella/bacteria’s rotational velocity (figure to right). This information along with the average wavelength of E. coli flagella, as determined by scientific literature, allow the program to compute the individual and average frequencies of flagella motion. The error of these calculations is also taken into account. The standard deviation of the individual frequencies is taken to quantify the error of the average response.


Our program creates a spatial representation of our data by plotting the individual and average frequency responses. The magnitude of the average response has been scaled to the number of individual responses it represents. This allows the viewer to get a better perception of the combined noise that these individual flagella would make, in other words: loudness. Furthermore, this spatial representation allows for a qualitative form of error analysis.


Another function of our program is to represent the average flagella frequency from an individual condition and create a musical representation. In order to place this the determined frequencies in a range that humans are more familiar with (C6 range), a scaling factor is used. From this, the program displays and plays a short sample of the corresponding note. This information is further used in our bacteria keyboard.