Bitbucket - I use this more than GitHub.
A web-based application that stores genomic variant information for over 3,800 patient's observed as Children's Mercy Hospital. It integrates many data sources; dbSNP, Exac and others. This tool also enables analysts to classify variants using ACMG guidelines.
A variant analysis tool, built in Java, that provides real-time filtering of next-generation sequencing data. It includes multi-sample analysis, mendelian inheritance analysis, variant filtering and many more features.
Pharmacogenetics is the study of genetic differences in drug metabolic pathways which can impact response to drugs; both therapeutic effects and adverse events. Storing an individual’s complete DNA sequence is costly, therefore, we store variants as haplotype sets. Haplotypes are variants occurring together on inheritable structures called chromosomes and are annotated in reference to their impact on protein production, pathway function, and potential to cause harm. Making use of haplotypes for collaboration amongst different research fields, labs and projects is difficult due to the lack of a formalized system. Many dispersed and disorganized attempts at maintaining haplotype sets have been attempted, but a single point of information storage for querying does not yet exist. By building a universal translation layer between the widely accepted human genome reference and single gene reference sequences, Astraea enables researchers to share, manage and provide custom meanings to their own nomenclatures easily through a web interface. The nomenclatures represent a compressed form of the variants across genomic regions saving costs in associated storage. The future goal of the application is to track drug reactions associated with appropriate genomic variants.
This tool converts your 23 and Me raw data file into VCF format.
A fun project that predicts whether you will win or lose a ranked League of Legends game based on a predictive model that I created. The model was built using ranked games of players in the Challenger and Master tiers. An interesting statistic that was found during the creation of the model was that the accuracy increased when lower tiers were tested. The accuracy ranges from 96% to 89% - bronze to challenger respectively.
This is a League of Legends data analytics project that performed timeline and event analysis of challenger tier summoner matches. This required me to create a classification algorithm to determine if the match consisted of a baron throw. Additionally, I created a data pipeline that automates the collection of every summoner's match in the challenger tier for every region. The end result shows which region has the most baron throws in the region and which summoners in each region have the highest or lowest baron throw rate.