The Cancer Genomics Cloud Summer Symposium - 2021

Overview

The Cancer Genomics Cloud (CGC) is hosting our first annual Summer Symposium, featuring Seven Bridges expert speakers, on August 18th, from 12:00pm to 2:30pm EDT. In this symposium, the speakers will focus on three areas of cancer research: Epigenomics, Image Processing using Machine Learning, and Single-Cell Analysis. To register for this event, click here:

What to expect at the Summer Symposium 

This symposium features three sessions, beginning sequentially after the introductory remarks: 

Introduction: 12:00pm - 12:10pm EDT

Epigenomics: 12:10pm - 12:55pm EDT

Image Processing using Machine Learning: 1:00pm - 1:40pm EDT

Single-Cell Analysis: 1:45pm - 2:25pm EDT

Featured Talks of the CGC Summer Symposium 2021

Expert Speakers: Jeffrey Grover, Ph.D, and Nevena Vukojicic

12:10pm - 12:55pm EDT

Epigenomics is a broad field encompassing diverse processes affecting gene expression. DNA methylation, histone methylation and acetylation, and other phenomena affect the landscape of chromatin accessibility and modifications. These have downstream effects on gene expression, and can be important markers of disease. Epigenomics data is large, and the Cancer Genomics Cloud is an ideal platform to perform analyses at the scale needed for an epigenomics study. The CGC Public Apps Gallery includes several best practice epigenomics workflows based on the ENCODE consortium, that are optimized for running on the cloud, thereby ensuring scalability and confidence in the reproducibility of results.

In this talk, we will be demonstrating the CGC’s capabilities to perform ATACseq and ChIPseq data analysis. The CGC Public Apps Gallery now includes best practice workflows for these experiment types based on the ENCODE consortium. ATACseq, or Assay for Transposase-Accessible Chromatin using sequencing, is a sensitive method that allows for the determination of accessible genomic regions. Accessible chromatin is frequently associated with increased gene expression. ChIPseq combines chromatin immunoprecipitation with sequencing to determine precise loci where proteins of interest interact with DNA. Proteins of interest are often transcription factors or modified histones, allowing researchers to determine the association between these features and other genomic loci or phenomena.

Using the CGC’s integrated Data Cruncher computational notebooks, interactive analyses may also be performed on the outputs of computational workflows. We will be demonstrating the use of Data Cruncher for downstream analysis, as well as for combining multiple-omics data types in order to generate more comprehensive biological insights.

Meet The Experts

Jeffrey Grover, Ph.D

Genomics Scientist 

Jeff received his Ph.D. in molecular biology from the University of Arizona studying the interplay between genomic features like DNA methylation, small RNA accumulation, chromatin structure, and gene expression. At Seven Bridges, Jeff works across teams to expand our bioinformatics analysis offerings, support different programs, and contribute scientific expertise in computational genomics and molecular biology.

Nevena Vukojicic

Bioinformatics Analyst 

Prior to joining Seven Bridges, Nevena worked as a Research Trainee at the Faculty of Biology, Department for Biochemistry and Molecular Biology. Her focus was on the biochemical and molecular basis of host-pathogen interactions. She has also worked as an intern at the Institut Pasteur in Paris; as AMGEN scholarship holder; and at Vienna Biocenter. She holds M.Sc. in Fungal Biology from the Faculty of Biology, University of Belgrade, where she is currently enrolled in Ph.D studies.

Expert Speakers: Soner Koc and Ana Stankovic

1:00pm - 1:40pm EDT

With the rapid growth of multi-omics data integrations and possibilities, we are witnessing a rise in digital image acquisition and related technologies. Image comprehension by computer programs has become an attractive and active topic in the machine learning field and in application-specific studies. Medical image classification is one of the most important problems in the image recognition area, and it aims to classify medical images into different categories to help with disease diagnosis or further research. 

On the CGC, we have developed a collection of various deep-learning, preprocessing, and utility tools and workflows that enable researchers to perform image class prediction without performing any coding themselves. These tools and workflows include quality control, image format conversion, image preprocessing, organization of image classes, parallel training of multiple model configurations, and unlabeled image prediction. Together, they are streamed forward to enable easy integration between one another, to provide an easy and logical analysis flow.

By utilizing deep learning-based methods and transfer learning approach, these tools offer more general use classifiers that are not constricted to one dataset or data type. With computing and data capabilities of the CGC platform, and utilizing multiple GPU instances to speed-up model training process, we are enabling a massive speed and price improvement compared to manual writing and running model training scripts. This is a highly extensible infrastructure, which can be customized for a specific dataset, application or research focus. 

Meet the Experts

Soner Koc

Genomics Scientist 

Soner received his MS in Computer Science from Bilkent University. His master's thesis is entitled: Deep Convolutional Network for Tumor Bud Detection, and focused on colon cancer tumor and biomarker prediction, segmentation, and localization with deep convolutional neural networks. At Seven Bridges, Soner has previously worked as both a Software Engineer before transitioning to a Genomics Data Scientist for the Programs Team.

Ana Stankovic

Bioinformatics Analyst

Prior to joining Seven Bridges, Ana studied Molecular Biology at the Faculty of Biology, University of Belgrade. She worked as a Marine biologist and then for several years as a Web Developer. As a bioinformatician at Seven Bridges she found the balance between programming and biology in developing bioinformatics tools and workflows.

Expert Speakers: Manisha Ray, Ph.D, and Dalibor Veljkovic

1:45pm - 2:25pm EDT

The Cancer Genomics Cloud powered by Seven Bridges makes the analysis of large datasets accessible from any environment, which is critical as the volume of datasets continues to expand. This expansion of the size, complexity, location of datasets has been prominent in the field of single-cell genomics, which can present challenges for data analysis. In this workshop, we will demonstrate the suite of single-cell tools available on the CGC, from primary processing to trajectory analysis. We will show how the tools and computational resources available on the CGC provide a path for any scientist to explore complex biological questions in the rich datasets generated through single-cell genomics.

Meet the Experts

Manisha Ray, Ph.D. 

Senior Scientific Program Manager

Manisha Ray is the Program Manager of the Cancer Genomics Cloud at Seven Bridges. Prior to Seven Bridges, she worked extensively in the field of single-cell genomics, including in the launch of the first commercial product for automating single-cell capture and preparation. She received her Ph.D. in Biochemistry from UCSF on the role of proteases in cancer progression.

Dalibor Veljkovic, M.Sc

Bioinformatics Analyst 

Prior to joining Seven Bridges, Dalibor studied Computer Science at the Systems and Signals Department at School of Electrical Engineering, University of Belgrade, where he focused on machine learning and biomedical signal processing. After this, Dalibor was part of the team working on Personalized Machine Learning for Autism Therapy with the MIT Media lab. As a Bioinformatician at Seven Bridges, Dalibor focuses on tools and workflows for single-cell analysis.

FOR MORE INFORMATION…

To learn more about the CGC and its resources, please see our documentation. For any additional inquiries, please contact support@sbgenomics.com

Unravelling the mysteries of the human proteome: From tools to workflows and analyses

Overview

Seven Bridges is hosting a Proteomics Virtual Workshop along with our collaborators from the NCI Proteomic Data Commons (PDC), on April 21st, from 1 pm to 3 pm EDT. In this workshop, participants will learn how to use the FragPipe proteomics pipeline, developed by members of the Nesvizhskii Lab at the University of Michigan for CPTAC data processing. In addition, participants will learn how to interact with the CGC platform and launch tools on the cloud. This shotgun proteomics pipeline itself consists of the software tools MSFragger, Philosopher, and TMT-integrator, and others.

What to expect in this workshop

This workshop will feature a hands-on demonstration of navigating the PDC website, linking publicly-available PDC data to personal projects on the CGC, without the need to download or upload, finding files in the CGC Public Apps Gallery, running the shotgun proteomics pipeline, and finally viewing the results. Specifically, this pipeline is designed to process high-throughput data-dependent acquisition (DDA) tandem mass spectrometry data acquired from peptides labeled with TMT tags. The multiplexed labeling, as applied in the CPTAC program, allows for differential quantitation across multiple tumor/normal samples. 

Who is the target audience?

We encourage all within the proteomics research community to attend. For this workshop, some basic familiarity with running proteomics processing pipelines is assumed. There is no need to be an expert, but familiarity with proteomics data processing, running pipelines, and Docker will be beneficial. 

What is needed to attend?

The main focus of the workshop will be how to run pipelines on the CGC, thus an active CGC account is required. To create an account on the CGC and learn about available pilot funds, click here.

The Team