New features and a new Portal released at ASHG

The Knowledge Portal team is back at work after a fantastic week at the American Society of Human Genetics meeting. We had many great conversations with researchers at our exhibit booth and at the Broad Institute exhibit booth, where we had a couple of guest spots. This year, we also held a workshop session on the Knowledge Portal Network and the Diabetes Epigenome Atlas (DGA), and about 80 people came to learn the basics of navigating the Knowledge Portals and the DGA. We were asked to provide the slides from that session, and they can be viewed here, but please note that they may not be easy to interpret without the accompanying oral presentation. We are working on creating both instructional webinars and short videos explaining different aspects of the Portals; stay tuned! And in the meantime, please contact us with any questions--we're here to help.

Part of the Knowledge Portal Network team at our ASHG booth

As usual, we released a number of new features on the Type 2 Diabetes Knowledge Portal in time for the ASHG meeting:

Calculated credible sets

Credible sets are useful because they assign to individual variants in a locus a probability of being causal for a phenotype. On Gene Pages (see an example), when viewing the type 2 diabetes phenotype, the Credible sets tab displays credible sets generated and published by Mahajan et al. (2018). However, credible sets have not been generated by researchers for phenotypes in the T2DKP other than T2D.

Now, the T2DKP provides calculated credible sets for all phenotypes. When viewing a phenotype other than T2D on the Gene page, the Credible sets tab is replaced by a Calculated credible set tab. This LocusZoom module, developed by our AMP T2D partners at the University of Michigan, automatically calculates posterior probabilities from p-values. Calculated credible sets include up to 10 variants; the credible interval covered by the set may vary, depending on the strength of associations across the region.

UK Biobank PheWAS

Recently, we added to the T2DKP another LocusZoom module for displaying phenome-wide associations. The PheWAS display, showing associations for a variant across all of the phenotypes included in the T2DKP, is the default visualization in the "Associations at a glance" section of Variant pages (see an example). Now, by checking the "Use UKBB data" box, you can view associations for a variant across about 1,400 UK Biobank phenotypes from an analysis performed by our AMP T2D partners at the University of Michigan.

New LocusZoom visualization shows variant associations across UK Biobank phenotypes

Forest plot visualization of variant associations

We also provide yet another LocusZoom visualization on a separate tab of the "Associations at a glance" section of the Variant page. The Forest plot is an alternative way to visualize phenotypic associations for a variant. In addition to displaying the significance of associations, the Forest plot also shows the direction of effect and the confidence interval for variant associations.

Forest plot on the Variant page

Genetic Risk Score module

The T2DKP now includes an initial version of the Genetic Risk Score module.  This is an instantiation of the same custom burden test that is found on Gene pages, but instead of using as input a set of variants across a gene, the module uses a set of 243 variants identified by Mahajan et al. (2018) that are significantly associated with T2D risk. The module draws on 9 different datasets, including 3 housed at the Broad Data Coordinating Center and 6 housed at the T2DKP Federated node at EBI. Just like the burden test, it allows you to choose a phenotype for analysis, adjust the set of variants if desired, filter the sample set by many criteria, and set custom covariates before running the analysis. The results obtained from this module can potentially reveal genetic relationships between phenotypes. The module is still under development, and we would appreciate your feedback on it!

New Knowledge Portal added to the network

At the ASHG meeting we unveiled the newest member of the Knowledge Portal Network: the Sleep Disorder Knowledge Portal (SDKP),  for the genetics of sleep and circadian traits. There is currently one dataset for sleep genetic associations in the SDKP, "UK Biobank Sleep Traits GWAS," which includes chronotype, sleep duration, insomnia, daytime sleepiness, and nap traits. Additional association datasets are available for type 2 diabetes and glycemic traits, anthropometric traits, measures of kidney function, and psychiatric traits, and more sleep data will be added soon.

Join the Knowledge Portal Network team!

At the Knowledge Portal Network (currently consisting of the Type 2 Diabetes, Cerebrovascular Disease, and Cardiovascular Disease Knowledge Portals), we are looking for energetic, talented people to help us produce web portals that aggregate and serve genetic association results to the world in order to spark insights into complex diseases. There are positions open for a software engineer to help in developing and producing these web portals, and for a technical release manager to manage and coordinate tasks during production and maintenance of the portals.

The positions are located at the Broad Institute in Cambridge, MA, a dynamic and exciting work environment where cutting-edge science is applied to critical biomedical problems.

Find more details and apply for the software engineer or technical release manager positions at the Broad Careers site.

Announcing the Cardiovascular Disease Knowledge Portal

We are pleased to announce the launch of the Cardiovascular Disease Knowledge Portal (CVDKP). Our collaboration with Dr. Patrick Ellinor, Dr. Sek Kathiresan, and their colleagues in the Atrial Fibrillation, Global Lipids Genetics, Myocardial Infarction Genetics, and CARDIoGRAMPlusC4D consortia has created a resource that offers world-wide open access to genetic and genomic information about atrial fibrillation, myocardial infarction, and related traits, with the goal of democratizing access to genomic data and accelerating cardiovascular genomics research.


CVDKP home page

The CVDKP is constructed on a software architecture originally developed for the Type 2 Diabetes Knowledge Portal (T2DKP), which is the central product of the Accelerating Medicines Partnership in Type 2 Diabetes (AMP T2D). AMP T2D is a public-private partnership between the National Institutes of Health, the U.S. Food and Drug Administration, biopharmaceutical companies, and non-profit organizations that is managed through the Foundation for the NIH. AMP seeks to harness collective capabilities, scale, and resources toward improving current efforts to develop new therapies for complex, heterogeneous diseases.

The ultimate goal of AMP T2D is to increase the number of new diagnostics and therapies for patients while reducing the time and cost of developing them, by jointly identifying and validating promising biological targets for type 2 diabetes. The T2DKP furthers that goal by aggregating, harmonizing, and displaying genetic association and epigenomic results along with user-friendly analysis tools, allowing research biologists who would not individually be able to amass and manipulate these large datasets to glean insights from the data.

We are working towards these same goals for other complex diseases, by extending the platform and analysis tools constructed for the T2DKP. In partnership with the International Stroke Genetics Consortium, we recently created a Knowledge Portal for cerebrovascular disease (CDKP) based on the same infrastructure. Now, with the advent of the Cardiovascular Disease Knowledge Portal, we have a three-member Knowledge Portal Network for the genetics of cardiometabolic and cerebrovascular disease.

Data in the CVDKP directly relevant to heart disease include genetic associations with atrial fibrillation, electrocardiogram traits, plasma lipid levels, and myocardial infarction. Additional association datasets are available for type 2 diabetes and glycemic traits, anthropometric traits, measures of kidney function, and psychiatric traits. You may browse the complete list of datasets and their descriptions on the CVDKP Data page.


As for the Cerebrovascular Disease Knowledge Portal, in the CVDKP we also continue to work with the American Heart Association Precision Medicine Platform (PMP) to provide an additional avenue for accessing cardiovascular genetic data. Currently, summary statistics from the AFGen GWAS and AFGen exome chip analysis datasets are deposited in the PMP.


We welcome all suggestions, comments, questions, and submission of relevant datasets for the CVDKP. Please contact us at help@cvdgenetics.org!