FINAL PROJECT PROPOSAL: Magnus Opus and Exigence > Predicting Synthetic Lethality: Machine Learning on Graph-vector embedding to Predicting Genetic Interaction Synthetic Lethality.
I will love reading this. Your choice about format. Do you have data displays with this work? Or conceptual visualizations?
Glad this class helps you with a real project. :)
Glad this class helps you with a real project. :)
November 29, 2018 |
Marybeth Shea
Yes, there are visuals and data displays
November 30, 2018 |
CS
Context: A research paper on predicting synthetic lethality for gene interactions and my contributions to the paper. Here’s a snippet of the intro:
Genetic interactions are measurements of relationships between genes. For this project, we will be looking to predict synthetic lethal gene interactions (SL) and non-synthetic lethal gene interactions (non-SL). A synthetic lethal interaction occurs between two genes when the perturbation of either gene alone is viable but the perturbation of both genes simultaneously results in the loss of viability (O’Neil, 2017). This means that two genes share a SL relationship if when a combination of deficiencies in the expression of two or more genes leads to cell death, whereas a deficiency in only one of these genes does not. Otherwise, an interactions between two genes is concerned to be a Non-SL.
…
Through machine learning on both individual species networks and merged species networks, we plan to find out if merging networks of two different species allow for better prediction of gene interaction. In essence, would merging the genetic makeup of two different species allow us to better understand the gene interaction within the species? Our end goal is to make a baseline for merging networks of different species, hopefully to give some insight to future projects in this domain
Purpose: Predicting SLs is important because SLs are a mechanism we can use to cause targeted cancer cell death, synthetic lethal genetic interactions with tumour-specific mutations may be exploited to develop anticancer therapeutics. One problem, however, is that predicting SLs can be hard to measure in humans. However, SLs can be measured in model organisms like yeast. In this paper, we look at two model species of yeast, Schizosaccharomyces pombe (Sp) and Saccharomyces cerevisiae (Sc). For both species we are given a network of genetic interactions and each category for each interaction, SL and non-SL. The hope is to predict genetic interactions among species accurately.
DOCUMENT TYPE: Research paper.
Design/format: TBA. The final report will be written in LaTex, in a traditional two column style. However, this will be merely be a (fairly premature) draft of the paper, so I may just turn it in a usual double spaces paper.
Citation style: APA for this class’s version.