Project: Analysis of Tumor Methylation Quantitative Trait Loci (tumor meQTLs) in Pediatric Cancer Patients with Central Nervous System Tumors

Project: Predicting the Subcellular Localization of RNA-Binding Proteins• Built a classifier for inferring whether proteins bind to RNA in the nucleus or outside it• Derived features from eCLIP data including whether binding sites overlap an exon, are of a spliced or unspliced context, span an exon-exon border, overlap a 3’UTR, and whether sites are from a protein-coding transcript• Worked full-time over the summer; Working part-time, remotely during the school year to write manuscript

• 1 of 30 students selected for the Ohio State Office of Student Life’s premier leadership development experience• 2-year interdisciplinary program; Weekly professional & leadership development sessions• Work with company or nonprofit on pitch challenge each semester• Traveled to Switzerland and Italy in March; Met industry/research leaders at CERN, Mettler-Toledo International, & FIFA; explored Zurich, Geneva, & Milan’s history and culture; Traveled to Rhode Island in September• Moderated panel on the age of AI at a student-led event last year; Volunteer at Columbus community center

Project: Analysis of Tumor Methylation Quantitative Trait Loci (tumor meQTLs) in Pediatric Cancer Patients with Central Nervous System Tumors• Developed project plan for undergrad thesis using quantitative trait loci framework similar to previous summer project• Gained access to data for 642 patients through the Molecular Characterization Initiative for Childhood Cancers• Processed, filtered (QC), and normalized methylation data and computed genotype principal components covariates• Preliminarily mapped significant tumor meQTLs using a linear regression-based approach• Plan to colocalize meQTLs with germline cancer predisposition variants, mutations in transcription factor binding motifs, and CTCF-binding sites, as well as explore genes with transcription start sites close to associated CpG sites

Project: Analysis of Histone Modification Quantitative Trait Loci in Opioid-Overdose and Control Brain Tissues• Mapped quantitative trait loci using ChIP-seq and genotype data from post-mortem nucleus accumbens tissue• Corroborated findings with alternative mapping approach by determining correlation between effect size outputs• Performed opioid overdose case-specific quantitative trait loci mapping; Found sites uniquely significant in this case

Project: Influence of Media Coverage on The U.S. Supreme Court Docket• Interviewed and assigned students to text analytics roles within 4 political science research teams• Led Supreme Court project investigating relationship between granted writs of certiorari & degree of news coverage• Wrote Python scripts to scrape information for each case recorded in the Supreme Court Journal, including case name, decision year, court of origin, state of origin, and outcome of petition• Ran query using LexisNexis’ API, pulling mentions of each case in news articles prior to the Supreme Court granting or denying review over the past 15 years; Parsed resulting JSON files to get media mentions value for each case

Project: Classifying Pathogenic Copy Number Variants (CNVs) with Machine Learning• Validated two classifiers with differing training data stringencies using scraped testing data from DECIPHER• Explored potential model features and implemented/assessed existing CNV significance classifiersProject: Genome Aligner Comparison• Compared accuracy/speed of 3 tools (bwa-mem, bwa-mem2, dragmap) using set of benchmark reference samples• Used AWS Step Functions to produce VCF files from BAM files; benchmarked variant calls against truth dataset• Also worked on manuscript for ancestry project during this semester: produced confusion matrix, ROC curve, and precision-recall curve figures and researched/summarized existing methods for ancestry inference

Project: Characterizing Low Complexity Regions (LCRs) in the Salmonella Typhimurium Proteome• Called LCRs using tool that produces self-comparison dot plots then runs kernel convolution and segmentation• Queried meta-server for LCR annotations to validate approach and computed residue/type compositions within LCRs• Characterized LCR prevalence within cell structure categories and among virulence proteins

Project: Comparison of Variant Effect Prediction Tools• Compared models for predicting effect of somatic variants, particularly those incorporating protein structure features• Classified deleterious variants using genome to protein coordinate mapping and a database of mutated amino acidsProject: A Machine Learning Approach for Classifying Somatic RNA-Seq Variants • Visualized RNA variant data with PCA, t-SNE, UMAP, and FAMD plots in Python and R • Built logistic regression and random forest models for classifying somatic variants in Python

Project: Impact of a Mediterranean Diet and BMI on Gut Viral Communities• Utilized bioinformatics software to classify viruses• Performed diversity analyses and statistical tests in R

Project: Algorithms for Inferring Ancestry from Genetic Variants • Built algorithms for ancestry inference in Python using dimension reduction techniques (PCA and UMAP) and machine learning