'Inspiring innovation through forums for idea-sharing'
CBIIT TechScouts is a forum for promoting continuous improvement across CBIIT through the cross-fertilization of ideas, experiences and recommendations. This forum is designed to foster new collaborations, learn about opportunities to better serve our customers, engage CBIIT more broadly, and raise awareness of new techniques and technologies that promote innovation. Central to the CBIIT Tech Scouts is the focus on gathering insight from the community with ideas and experiences on how emerging information can be used to improve scientific productivity and accelerate cancer research.
Table of Contents
Topics Archive
Date | Topic (CLICK FOR MORE INFO) | Topic ID | Summary |
---|---|---|---|
| Semiconductor Feature Size Roadmap | TS-0016 | Pretty amazing feats of science and engineering to get us to 2020. https://en.wikipedia.org/wiki/5_nanometer For comparison, a single benzene ring is about 500 picometers, or 0.5 nanometers |
| Monarch Made Twitter | TS-0015 | I wanted to share a session that was presented at the AWS Public Sector Summit by NIAID that discussed Next-Generation Medical Analysis leveraging the cloud. Our CSRA team at NIAID assisted by developing a PaaS solution they’ve named Monarch, discussed in the presentation, that assists with this work leveraging a full DevOps pipeline. For background, Nephele, discussed in the presentation, is a project that a team in BCBB has been working on for several years now. Due to the infrastructure group (OEB) not having a defined service offering for public cloud at NIAID years ago when Nephele started, the project team went out on their own and architected a solution using AWS. Within the last year, CSRA consulted with them and provided some guidance for best practices and improvements now that we are launching a custom platform as a service using open source technologies and AWS services. In the future, we’ll be moving Nephele components to the custom PaaS we created called Monarch and hosting it under our architecture. https://www.youtube.com/watch?v=rLkBwWv0Hdc&list=PLhr1KZpdzukePsKIUofhgp50b63-5yr1V&index=66 |
| DCEG Linkage: 3D Printing in Radiation Research | TS-0014 | Here's a unique example of 3D printing right on site in the NCI Shady Grove building. This is from Choonsik Lee's group in DCEG's Radiation Epidemiology Branch. https://dceg.cancer.gov/news-events/research-news-highlights/2017/3d-printed-phantoms?cid=eb_govdel |
| SciAm: China Shatters 'Spooky Action at a Distance' Record, Preps for Quantum Internet | TS-0013 | Here's more on the long-term future (ignore the immediate or mid-term geopolitical implications). https://www.scientificamerican.com/article/china-shatters-ldquo-spooky-action-at-a-distance-rdquo-record-preps-for-quantum-internet/ |
| MaruOS - Your Phone is Your Computer | TS-0012 | With Maru, your phone is your PC -- you connect it to a keyboard and monitor whenever you need the desktop environment. Obviously this isn't ready for government use yet, but it illustrates the path toward convergence into single device personal computing. |
| Why Windows Must Die. For the Third Time | TS-0011 | Interesting article about the history—and potential future—of Windows: http://www.zdnet.com/article/why-windows-must-die-for-the-third-time/ Great article as Microsoft tries to stay relevant in an age of minimal OS’s, immutable servers and relentless competition from, some would argue, more mature operating systems such as those found on MAC’s. It brought a tear to my eye as I took a walk down memory lane |
| Biowulf HPC Expansion Details | TS-0010 | Biowulf now has >180,000 CPUs, including 100 GPU nodes, 24 very large memory nodes (up to 3TB per machine), and a very high-performance, low latency network (FDR infiniband) with an 80 Gb/sec connection to the outside world. Additional dedicated resources (with priority usage) are available to CCR researchers (10,000 cpus) and, soon, to DCEG (4000 cpus). This expansion will likely place Biowulf in the top 100 most performant HPC systems in the world. If you have questions, feel free to contact me or the Biowulf staff. |
| Globus Data Management Solution | TS-0001 |
|
| The Algorithm Will See You Now | TS-0009 | Great piece in the April 3 New Yorker on medical diagnostics and machine learning. They used Thrun and Hinton’s Columbia work with predicting melanoma as an example. http://www.newyorker.com/magazine/2017/04/03/ai-versus-md Anyone interested in machine learning – FAES offers python based machine learning in BIOF 509. Taught by Jonathan Street and Burke Squires on campus. |
| TEDxBuffalo - Become a Citizen Data Scientist | TS-0008 | Here's a TEDx talk about how normal (i.e. non-academic) people can be enlisted to support certain classes of scientific research. As the talk alludes, there are many possibilities for leveraging citizen science, and it is useful to think about the technologies available and the ways people interact with technology (like gaming or idle time-wasting on a cellphone) to understand those possibilities. The value of citizen science is not just in distributed data collection or analytical tasks that are currently easier for humans than computers. It is also the engagement of more people in the purpose, goals and problems of current research efforts. https://www.youtube.com/watch?v=zgijAGkdjZc&feature=youtu.be |
| Oracle Doubling the Cost to Run in AWS | TS-0007 | Found these two articles related to Oracle’s new pricing strategy which effectively doubles the cost to run Oracle products in AWS. For those of you that have worked with Oracle over the years, I’m sure this isn’t a big surprise, but as we move to the cloud it may be a driver to move to other database platforms such as AWS RDS or Red Shift on AWS. Of course I suppose there’s always the Oracle cloud. https://www.cloudtp.com/doppler/oracle-is-now-charging-double-to-run-on-aws/ https://www.theregister.co.uk/2017/01/30/oracle_effectively_doubles_licence_fees_to_run_in_aws/ |
| IBM Watson for Oncology | TS-0006 | Watson provides clinicians with evidence-based treatment options based on expert training by MSK physicians.Whether a community oncology practice or an international hospital, oncologists like all clinicians are struggling to keep up with the large volume of research, medical records, and clinical trials. Watson scales vital knowledge and helps oncologists. Now, with the collaboration between IBM and MSK, Watson for Oncology utilizes world-renowned MSK expertise to evaluate specific details of each unique patient against clinical evidence. |
| Data Science Job Report: R, SAS, Python | TS-0006 | For anyone paying attention to technology trends in statistical analysis and data science, here's a report from an R-focused data science blog: http://r4stats.com/2017/02/28/r-passes-sas/ The methodology used to track these trends involves data from job postings. Specific trends within cancer research are likely to be somewhat different. |
| DNA - the Future of Data Storage? | TS-0005 | You've probably heard about the concept of storing data in DNA by now. Here's a quick what/why/how/when from Yaniv Erlich's lab at Columbia U: https://www.researchgate.net/blog/post/dna-could-be-the-future-of-data-storage |
| NCI Genomic Data Commons R Client | TS-0004 | The NCI has established the NCI Genomic Data Commons to store and distribute NCI datasets. One powerful feature of this new resource is a RESTful API. The Bioconductor project is a large, open-source for the analysis and comprehension of -omics datasets. A couple of us are working on an Bioconductor package for finding and accessing data in the NCI Genomic Data Commons directly from R, exposing a community of about 4000 developers and 100,000 users to NCI genomics datasets. The GenomicDataCommons package is available on GitHub and will be making its way into the Bioconductor project in the next month or so. |
| Open, Reproducible Neuroscience | TS-0003 | Towards open and reproducible neuroscience in the age of big data Abstract: The Stanford Center for Reproducible Neuroscience (CRN) was founded to develop a number of initiatives to help to make neuroscience more open and reproducible. These initiatives include influencing the scientific culture (e.g. The OHBM Replication Award), introducing new standards for organizing neuroimaging data (BIDS) and software (BIDS Apps) as well as building tools to assess MRI quality (MRIQC) and perform robust preprocessing of fMRI data (FMRIPREP). Tying all of those efforts together, as well as capitalizing on the experience gained from OpenfMRI and NeuroVault, is the upcoming OpenNeuro project. This novel software-as-a-service platform seeks to harness the power of high-performance computers to provide free data analysis in exchange for making data publicly available. There is a nice technology feature article in Nature on "big data" in Neuroscience, touching on many of those parallels -- data management when dealing with 100s of terabytes, attitudes toward data sharing, and the notion that in many cases "big data" is still not enough data. Neuroscience: Big brain, big data http://www.nature.com/nature/journal/v541/n7638/full/541559a.html |
| AWS CloudTrail | TS-002 | Monitoring changes to infrastructure and data is a very important part of a robust and secure infrastructure. This service allows ongoing logging of API calls (so, changes) to infrastructure and even supports S3 events (for changes to data). https://aws.amazon.com/cloudtrail/ Other cloud providers offer similar services directly or via third-parties. |
// |
Show and Tell Archive
Date | Topic | Who | Notes | Action Items |
---|---|---|---|---|
Attachments
Welcome to CBIIT TechScouts!
We are a forum for promoting continuous improvement across CBIIT through the cross-fertilization of ideas, experiences and recommendations.
The goal of CBIIT TechScouts is to collectively improve scientific productivity through the sharing of knowledge.
What We Do
CBIIT TechScouts is designed to foster new collaborations, learn about opportunities to better serve our customers, engage CBIIT more broadly, and raise awareness of new techniques and technologies that promote innovation.
Central to the CBIIT Tech Scouts is the focus on gathering insight from the community with ideas and experiences on how emerging information, data-oriented, and software technologies can be used to improve scientific productivity and accelerate cancer research.
Our Service Includes:
- An email distribution list, allowing members to stay updated on the latest topics and trends
- To join: email Miles Kimbrough, subject 'TechScouts Access Request'
- To join: email Miles Kimbrough, subject 'TechScouts Access Request'
- An annotated archive of topics and presentations, organized for future reference
- A monthly summary found on CBIIT Central, providing a lightweight overview of recent trends
- And finally, Monthly Show and Tell Sessions, opening the door for new ideas to be presented both in person and virtually
- To present: email Miles Kimbrough with topic and availability
Questions?
General Support : Miles Kimbrough | miles.kimbrough@nih.gov | 240.276.5251
Consultation and Guidance : Eric Stahlberg | eric.stahlberg@nih.gov | 240.276.6729
Technical Support : George Zaki | george.zaki@nih.gov | 240.276.5171