Bioinformatics Training & Education Program

If you’re interested in learning how machine learning can help you tap into the rich data within electronic health records, join us to learn more about the “Memorial Sloan Kettering (MSK) Clinicogenomic Harmonized Oncologic Real-World Data set” otherwise known as MSK-CHORD. It includes data for over 25,000 cancer patients to help you identify clinicogenomic relationships that aren’t as obvious in smaller, siloed data sets.

Dr. Nikolaus Schultz, an NCI grantee and computational oncologist at MSK Cancer Center, will give an overview of the data set and demonstrate:

• the feasibility of MSK-CHORD’s automated annotation from unstructured notes.
• how MSK-CHORD can train machine learning models to predict patient outcomes.

Some of the resulting data from studies leveraging MSK-CHORD are available via cBioPortal.

This talk will focus on analyses of the Patient Outcomes Repository for Treatment (PORT) which is a large registry of chronic pain treatment outcomes from patients seen in the pain clinics at the University Pittsburgh Medical Center (UPMC). Using methods such as propensity scoring, stratified modeling, and supervised machine learning, we can determine which treatments for chronic pain are or are not effective, the phenotypes most responsive to each treatment, and predict which treatments will be most effective in any new patient based on their phenotype (such as medications, injections, physical therapy, or mental health care).

This one and a half-hour online training covers the basic principles of FAIR (Findable, Accessible, Interoperable, Reusable) data and why it is important to make your data FAIR.  

 By the end of this training, attendees will be able to:  

• Define FAIR data   

• Explain what purpose FAIR data serves 

• Apply FAIR data principles to make data findable, accessible, interoperable, and reusable 

This is an introductory level training. 

Abstract: Spatial transcriptomics (ST) is evolving rapidly as a pivotal technology in studying the biology of tumors and their associated tumor microenvironments (TME). However, where/how it can impact research remain unclear. In this talk, Dr. Chen will describe their recent study that performs rigorous benchmarking amongst the single cell ST platforms CosMx, MERFISH, and Xenium (uni/multi-modal) platforms (PMID: 41006245) and several computational approaches that performs multimodal ST data fusion, identifies locations and directions of spatial transcriptomic gradients (STG)  (PMID: 38562886) and infers metabolic flux (PMID: 37573313) from ST data. Finally, he will describe their experience in studying factors deriving immune checkpoint therapy resistance in HPV+ H&N cancer patients.

Bio: Dr. Chen obtained B. Eng. from Tsinghua University (Beijing) and Ph.D. in Electrical and Computer Engineering from University of Illinois at Urbana-Champaign. He is currently a full professor in the department of Bioinformatics and Computational Biology at the University of Texas MD Anderson cancer center. His primary interest is to develop computational methods to analyze and interpret high-throughput human genetic, functional and clinical data towards understanding the evolution of cancer as a consequence of genetics and environment and identifying molecular targets useful for cancer diagnosis and therapeutics. Among the computational tools he developed, BreakDancer, VarScan and Monovar have been widely used for characterizing genomes and transcriptomes of tumor tissues and single cells.

In this lesson, attendees will learn the basics of ggplot2 to create simple, pretty, and effective figures with R. 

In this lesson, attendees will continue learning how to create publishable figures with ggplot2. Topics will include statistical transformations, coordinate systems, and themes. 

In this lesson, attendees and instructor will work together to craft a publishable volcano plot using the skills previously learned. 

Gil Kanfer, PhD, of the NCI CCR High-Throughput Imaging Facility (HiTIF), in the Laboratory of Receptor Biology and Gene Expression (LRBGE), will present the spatial biology analysis stack HiTIF is building to support Center for Cancer Research (CCR) researchers, with a focus on high-resolution spatial transcriptomics and multiplex protein imaging platforms existing in CCR Cores (e.g., Visium HD, Xenium-5k, CODEX) and how they can be turned into robust, reusable analysis workflows.

Using recent liver cancer and melanoma projects run by CCR investigators with the NCI CCR Single Cell Analysis Facility (SCAF) and Spatial Imaging Technology Resource (SpITR) core facilities in collaboration with HiTIF as examples, he will show how in-house algorithms—such as a zonation prediction model for mapping periportal vs pericentral regions and custom methods for collagen-based proximity and niche analysis—are combined with open-source tools to align images, integrate RNA and protein data, quantify cell-type composition and spatial organization, and systematically screen ligand–receptor interactions across conditions and time points. The talk will emphasize generalizable, technology-driven pipelines that take core-generated images all the way to quantitative, biologically interpretable spatial-omics readouts for CCR labs.

Attendance at this event is limited to NCI CCR personnel.

This lesson introduces general recommendations and tips to consider when creating effective and reproducible visualizations. Additional topics to be discussed include multi-figure panels, complementary or related R packages, and the use of ggplot2 in functions.