Events

Augmented Intelligence for Healthcare: How Can NLP Help Physicians at the Bedside?

Dr. Yanjun Gao is a postdoc research associate in the Critical Care Medicine (ICU) Data Science Lab in the Division of Allergy, Pulmonary and Critical Care Medicine within the Department of Medicine, at the University of Wisconsin Madison (UW-Madison).

Cross-language Data Development with Apache Arrow

Tired of running into data development challenges like conversion, datatyping, and performance without a unified standard? This talk will cover Apache Arrow: a unique opportunity to enable integrated multi-lingual and high-performance data development. We’ll cover a brief background, how Arrow compares to similar software, and some development examples.

Bioinformatics Algorithm Development for Mass Spectrometry Data and Its Application to Precision Medicine

Dr. Shen received his Ph.D. in Bioinformatics and computational metabolomics from the University of Chinese Academy of Sciences. His overarching research mainly focuses on bioinformatics algorithms/software development for multi-omics data and their application to precision medicine.

Sex Differences at the Molecular Level in Health and Disease: Lessons from the Human Transcriptome

The overriding research interest of the Stranger Lab is trying to disentangle the relationship between genotype and phenotype, and to understand the forces shaping functional genetic variation in humans. The lab combines experimental and computational approaches, such as high-throughput functional genomics and genome-wide association analysis (GWAS), to identify variable regions of the genome with functional effects on molecular traits, such as the transcriptome and proteome, and to investigate the context-specificity of these relationships and their role in human disease. The Stranger lab’s examination of sex differences in gene expression as part of the GTEx project revealed that 37% of all genes exhibited sex-biased expression in at least one GTEx-assayed tissue and identified gene-trait associations driven by sex-specific gene expression.

Improving the Reliability of Real-World Evidence Generation from Large-Scale Observational Data: Applications to Effect Estimation and Health Equity

Linying Zhang is a doctoral candidate at the Department of Biomedical Informatics at Columbia University. Her work centers on causal machine learning methods for reliable evidence generation from observational data.

Topic Alignment Identifies Functionally Relevant Vaginal Microbiota Sub-Communities

Laura Symul obtained her Ph D. in computational biology from the École Polytech-nique Fédérale de Lausanne (EPFL) in Switzerland, where she studied the molecular regulation of the circadian clock. She combined analyses of -omics data with mathematical models of the regulatory dynamics to infer quantities, such as mRNA degradation rates, that are otherwise not measurable.

Temporal Machine Learning for Phenotype Modeling with EHR Data

As a social scientist, data scientist, and clinical informaticist, his most recent work has focused on architecting visual analytics application to explore data quality in Electronic Health Records data and characterize patients, using Statistical Learning techniques and Data Science methodologies.

Artificial Intelligence and Imaging for Personalized Medicine

Dr. Jiang’s work focuses on the development and application of novel machine learning and deep learning approaches for medical imaging analysis and precision medicine. His work spans across multiple imaging domains and modalities including radiology as well as histopathology image data.

Addressing Frailty and Fall in Older Adults: Integrating Biomechanics, Wearables, and Machine Learning

Dr. Toosizadeh achieved his PhD in Industrial and System Engineering Department at Virginia Tech in 2013 with a Human Factor focus. His research focus is computational model, sensor-based engineering approach, and machine learning tools to diagnose and treat older adults with aging-related conditions.

Short Tandem Repeat Expansions are Under-Appreciated in Rare Disease Diagnosis

Dr. Dashnow completed her PhD with Alicia Oshlack at the Murdoch Children's Research Institute and the University of Melbourne, Australia.

Unsupervised Statistical Methods and Data-Driven Analysis Workflows for Spatially-Resolved Transcriptomics

Dr. Lukas Weber is a postdoctoral fellow at Johns Hopkins Bloomberg School of Public Health, and his research is on the development of unsupervised statistical methodology and software to extract biological insights from high-throughput genomic data.

Everything You Need to Know About Qualtrics but Did Not Know to Ask

Qualtrics is a powerful platform for forms and surveys; it is free to all CU faculty and staff upon request to OIT. The key features that set Qualtrics apart are complex skip logic and piping data to and from external databases. Other features include visual configurability, versioning, respondent mail merge, respondent tracking/reminders, confirmation emails, as well as numerous question formats.

Presented by: Precision Health and Genomics: Indigenous Mentoring and Ethics (PrIME) Supported by: Dept. of Biomedical Informatics, NIH NIGMS, Human Medical Genetics and Genomics graduate program Hosted by: Dr. Katrina Claw

Mathematical Modeling of Metabolic Dynamics in Adolescent Girls

Dr. Behn earned her PhD at Boston University and completed postdoctoral fellowships at Harvard Medical School and the University of Michigan. Her research is at the interface of mathematical modeling, data science, and physiology and focuses on scientific and translational questions in sleep/wake behavior, circadian rhythms, whole body metabolism, and interactions among these systems.

Technology, Humans, and Healthcare

Presenting three topics related to human-centered design in a clinical environ-ment. The first topic is the importance of delivering usable guidelines that clinicians can understand and apply in their practice. The second topic is the importance of involving clinicians in the user-centered design process. The third topic is the importance of working with individual clinical experts.

Three cool projects, three key ideas

In this talk, I will pick three projects from our group and use them as the backdrop to describe three key ideas in supervised machine learning (ML): i) Learning patterns in large molecular networks, ii) Turning text into numbers, and iii) Predicting things about biological samples based on the data collected from them. I will end with a discussion of the importance of being really careful in evaluating the performance of ML models.