Title: Data Analysis and Software Design to Assist Researchers in Choosing Effective Endogenous Genes for CRISPRa
CRISPR activation (CRISPRa) is a tool used in synthetic biology to activate genes. However, there are stringent rules for where CRISPRa can effectively promote transcription. Many of these rules are either unknown or have not been compiled into a single model. In order to rectify this, I worked on two main software projects. (1) Mock Data Generation Model for FACS-seq CRISPRa was a project where synthetic data was generated in the style of a CRISPRa experiment. (2) P. putida RNA Seq Activation Filtering was a project where the goal was to find effective candidates for CRISPRa in Pseudomonas putida’s endogenous genome. Click read more to get a more thorough description.
In the summer of 2021 I had the privilege of working as an R&D intern focused on data science at Sandia’s Center for Cyber Defenders. I joined a team engaged in vulnerability assessment of software. For my summer project, I was asked to research the efficacy of data analytics as a means of expediting software behavior analysis. Early results showed that NLP-based techniques developed; such as n-gram divergence comparisons and UMAP dimension reductionality; successfully differentiate event logs collected under varying conditions.
Capstone Project.June 17, 2021 By Joely Nelson, Kenny Krivanek, Joe Ammatelli, and Tevin Stanley
The Vaccine Impact Modeling Tool is an interactive modeling tool, given user input, retrieves COVID related data, models pandemic outcomes, and displays them on a global map.
Made as the final project for CSE 482B: Capstone Software Design to Empower Underserved Populations at the University of Washington.
A presentation of some of the security vulnerabilities in machine learning. I cover model extraction attacks, model inversion attacks, and adversarial example attacks.
Made as the final project for CSE 584: Computer Security at the University of Washington.
Deep Learning Project.Dec 14, 2020 By Joely Nelson
In this project, I developed a convolutional neural network (CNN) to classify DNA sequences from two data sets. I mimic the architecture of the CNN used in prior work on two different datasets, and achieve close to the paper’s accuracy.
Data Science Research.May 15, 2020 By Joely Nelson
CRISPRa is a genetic engineering technique that allows for a flexible engineering of metabolic pathways which have applications in medicines, biosensors, and much more. However, there are many unknowns, and much of CRISPRa engineering has been seeing what works with trial-and-error.
In my research in the Carother’s Group, I have worked on taking a data driven approach to CRISPRa engineering to reduce this trial and error. By using machine learning, embedded system design, and analytics on genetic datasets, I was able to find actionable insights and find areas for further study.
Web Application.Mar 12, 2020 By Joely Nelson, Kyle Johnson, and Robel Wondimu
An interactive web application designed to visualize presidential voting data from prior years, as well as display polling trends from 2020. In this project I was involved in the design, extensive data cleaning, and some of the JavaScript programming.
This application can be viewed at https://uw-cse442-wi20.github.io/FP-prediction-2020/. Please use Google Chrome.
Embedded System.Mar 6, 2020 By Joely Nelson, Ian Davison, Matthew Trahms
I was heavily involved in developing the software for a simplified battery management system.
My main work on the system was implementing and designing the touch screen functionality and informative display. I also implemented safety interrupts to open contractors based on changing voltage, current. I was also responsible for writing code to use interpolation and lookup tables to calculate state of charge from voltage and temperature.
Made in collaboration with Ian Davidson and Matthew Trahms for CSE 474.
Embedded System.Aug 21, 2019 By Joely Nelson Associated with the Carother’s Research Group
It was proposed that using a blue light to control the induction of molecules in the CRISPRa process could make data collection for CRISPRa experiments easier, less error prone, and more reproducible.
I designed a blue light projector controlled by the Arduino Uno.
I was able to create a simple projector. I had no prior experience with circuits or the Arduino microcontroller, but was able to create this project.
A preliminary experiment found that the blue light has promise in being able to induce the expression of molecules. Further experimentation is needed.
Python Package.December 13, 2017 By Joely Nelson and Eric Klavins
A python framework that uses numpy and sypy to make it easy to quickly generate and analyze chemical reaction networks. This package was created by expanding off of Eric Klavins’s code for CSE 486.
It allows the user to model chemical reaction networks, automatically generate stoichiometric matrices, analyze the equilibrium, graph uncertainty, and perform other mathematical operations
With just a few lines of code a user can create a model of the central dogma of biology: RNA being generated, and that RNA becoming protein. Below is an example of this model. The x axis represents time, the y axis represents the amount, and the shaded areas represent the uncertainty and noise of what is predicted.
The user below has used Sbjiojn to model the same system above, but has used random state jumping to observe the system uncertainty.
Using data from WDFW, WSDOT, and Defenders of Wildlife, I used ArcGIS to determine the location of barriers to wolf spread to the Pacific Coast Region, and propose road sections where the construction of wildlife corridors could help wolf dispersal into this region.
I followed the assumptions of Richard Walker and Lance Craighead who created models to find places where connectivity could be increased in the Northern Rockies in Montana. These assumptions are:
Humans pose problems for successful transit
Good corridors are composed primarily of preferred habitat types
Current human developments are permanent.
Made for the final project of Geography 360 at the University of Washington.
