Computer Science & related.

File.parse() + Inventory.write()

The broader picture goal of this project was to automate the process of a polymerase chain reaction and minimize human lab error.

Most recently, I implemented a parser and inventory from scratch for a set of human created files pertaining to data of existing reactants and dnas inside of a laboratory. HashMaps are used to store attribute: value pairs for each tube in a box. After an experiment, the contents of the laboratory are changed, thus new human readable text files are generated with the write() method.

CellCounter

Aided in designing an algorithm to count the number of cells (blue pixel islands) that overlapped with blood vessel stain (red). To remove low intensity background, we easily set a threshold that rejected most of it. However, often the highest intensity “blobs” in an image were not true cells, but rather artifacts of assay imprecision. To remove the high intensity background, I suggested filtering the image for eccentricity: the ratio of radial lengths of a “blob”. Specifically, we rejected blobs whose eccentricities were significantly above the value expected for actual cells.

RNAfold.py

Above is my implementation of the RNAFold module on a riboswitch strand of RNA to simulate its secondary structure with specific OR- logic gate parameters.

Percolate.java

Below are two images taken while my implementation of percolate.java runs on a graphics model called Standard Draw. This function randomly “opens” white spaces until there is a path for “water” at the top to flow all the way through.

ParseR&J.py

This project involved parsing the famous Shakespeare tragedy Romeo and Juliet and using various data structures to aggregate the most common and least common words and inspect those elements. It turns out that the 10 most common words in the play represent the structure of the english language. The least common words demonstrate Shakespeare’s rich vocabulary.

Plotting word count word rank (its position in the frequency hierarchy), I observed an exponential relationship. Plotting log(word count) versus log(rank), the graph becomes more linear. This means, for example, that if “orange” is the 6th most common word, and “apple” is the 7th most common word, “apple” would likely show up about half as many times as “orange”.