griffithlab / civic-meeting Goto Github PK
View Code? Open in Web Editor NEWRepo for advertising and organizing CIViC hackathon/meeting activities
Repo for advertising and organizing CIViC hackathon/meeting activities
Some people from the Stanford VCI are coming and we would like to discuss technical details of how we could use the Evidence Registry (Baylor) to help VCI users leverage curation work previously done in CIViC to make final pathogenicity assertions in the VCI.
Work with NCI-Thesaurus to create entire WHO representations, similar to AML (https://nciterms.nci.nih.gov/ncitbrowser/ConceptReport.jsp?dictionary=NCI_Thesaurus&version=22.02d&code=C159157&ns=ncit)
Establish list of relationships between who entities and the ICD codes
Proposed by Beth Pitel. This topic was the last one tackled in the afternoon session. 8 participants attended this discussion.
This work is underway but it may still be in play when the meeting happens.
Related issue from civic-server repo: griffithlab/civic-server#171
Currently, clinically relevant fusions are under-represented in CIViC (and other knowledgebases) in relation to SNVs and other small mutations. Fusions represent a unique curation challenge in CIViC as they involve multiple genes and multiple (sometimes difficult to define) genomic coordinate sets. They also represent unique challenges for developing NGS sequencing panels and may require specific technologies (e.g., Archer) to identify them. However, many fusions have clear implications for prognosis, diagnosis, or drug response. I propose a working group to tackle the specific challenge of curating fusions in CIViC.
Interested to discuss how to motivate people to participate in curation efforts. Wall of shame (or tracking participation), rewards associated with achieving levels, points, reminders of contributions, small completable tasks, goal setting/goal completions, any thoughts to tie this into health apps like vitality etc -- i.e. companies pay for healthy workers, maybe interested to pay for healthy science participation too?
A recent publication made recommendations for the curation of non-coding variants.
https://genomemedicine.biomedcentral.com/articles/10.1186/s13073-022-01073-3
How do the germ-line recommendations apply to curation of somatic variants?
NGLY1 deficiency is an extreme congential rare disease. We built led by domain-experts feedback a disease knowledge network for this rare disease. This project aim at examining web applications to enable the community to contribute directly in the update and curation of this resource.
slides: https://docs.google.com/presentation/d/1KB6dK9GJj0XQpaTxEq_-EqiAS_jghd3X5y735l1y3nA/edit?usp=sharing
NGLY1 deficiency network: http://52.87.232.110:7474/browser/
Would be nice to have a feature where complex form's state (such as Add Evidence) were saved when a user exits a page, to be restored when they return.
Biomarkers can have clinical implications as diagnostic, prognostic and predictive biomarkers. All of these can be found and created in CIViC. Yet, pharmacogenomic (genomic predictors of drug toxicity) and predisposing (genomic predictors of disease probability) are additional biomarker categories that might be worthwile to discuss as an addition to the database.
Elevator pitch to be lead by Matt Brush? @mbrush would you be able to give a ~3 minute elevator pitch on this and if there is interest help to lead a session on use of SEPIO?
Given the scale of potential variation, to what extent can protein-specific models of functional impact be used as evidence?
I've been developing a method to predict functional changes of missense mutations using simulations of protein structure.
https://www.cell.com/biophysj/fulltext/S0006-3495(20)33203-3
In short, molecular dynamics (MD) simulations are used to define the characteristic "wobble" using an all atom representation of a variant's protein structure. By projecting this variant specific profiles into a lower dimensional space (PCA), we find that variants cluster by their contribution to divergent disease mechanisms, which allows us to define a decision boundary for a classification model. That is to say, we can use these projections of variant dynamics to train an AI classifier that predicts the impact of novel variants based on the MD results.
An interesting result came when we applied this method to predict venetoclax resistance in BCL2. Shown below is a projection of resistant variants (red), sensitive variants (blue), and a re-sensitized double variant (rescue - first G101V, second E152A) into PCA space. The "rescue" variant becomes more wildtype, as also shown with the predicted Ki.
https://www.sciencedirect.com/science/article/pii/S0010482521008544?via%3Dihub
The Ki's were predicted by quantifying the distance of a given variant from the wildtype, and our preliminary data suggests the degree of divergence from the wildtype dynamics correlates to the degree of functional change in other proteins as well (see the Biophysical Journal article linked above).
So... to what degree can these methods be used as evidence? Or on the other hand, to what degree can curated evidence be used to label/train these sorts of models?
A group could be formed to discuss curation best practices, data model improvements, adoption of ClinGen (MVLD) and FDA proposals, ontologies, etc.
Elevator pitch to be lead by Greg Stupp (Su lab) @stuppie
Interpret the potential interactions of variant genes using heat diffusion methodology.
Add initial heat of variants to corresponding nodes of a protein-protein interaction network, proportional to the oncogenicity metric value for each variant.
This will produce a subnetwork of the interaction network that is focused on the most relevant proteins/genes
Conjecture: this network could then be augmented via addition of drug-variant / drug-gene interaction links to suggest relevant therapies, one that may expose drug effects that affect multiple variants.
From Discussion:
Have a mechanism where an expert can suggest other experts to work on a specific gene or variant. Optionally specify a missing gene or variant.
e.g. Zenodo, bioCADDIE for citable records
Currently only a few evidence items have been added for hepatocellular carcinoma - disease specific evidence mining could help to expand the database for this disease.
Example: BRCA1 or BRCA2
Which criteria are used to decide wether a SNP is benign?
(How) do you report a SNV, which is likely to be benign, but without enough conclusive evidence?
Integrate the CIViC DB with genome browsers including igv.js, JBrowse, and possibly IGV. This integration could take 2 forms, not mutually exclusive
(1) Link out to CIViC from VCF and other variant track types (e.g. SnpDB) and provide variant details in a popup.
(2) When zoomed in to βgeneβ level, query the CIViC API for variants as genes come into view and display as a track. Clicking on a variant would query the API and bring up details in a popup.
The motivation is to give researchers browsing their own variants in the genome browser easy access to CIViC data. igv.js does a similar thing now with Cravat.
If there is interest in this proposal I could outline an implementation approach here, or save it for the hackathon.
One noted gap in many well-known knowledgebases is the coverage and searchability of copy number variants and fusions/structural variants. Many our our in-house tools are attempting to employ segment/coordinate based searches to link to relevant data/knowledge for clinical reporting of copy number variants. Fusions are another beast for us - each assay seems to annotate things a little differently. For example, are the fusion partners always listed 5' --> 3'? How do we document structural rearrangements (translocations/inversions/CNVs) that result in over/under expression by position effect (i.e. they don't result in a fusion product)?
Review innate/ acquired genomic resistance mechanisms to targeted agents, how to interpret them and present the results of our curation in a schematic way.
There is interest in integration of CRAVAT and CIViC (possibly bi-directional?). In what specific ways? To be discussed with the Karchin Lab (https://github.com/KarchinLab).
We should discuss the CIViCpy module and recipes for its use in querying data over the CIViC API.
I am planning to attend the upcoming meeting/jamboree and would like to propose backfilling/entering CIViC NLI selections if there is still room for proposals. This would involve curators selecting evidence sentences/spans in PMC/pumbed articles which are support for existing (or new if preferred) CIViC evidence items.
Secondary Task: Labelling pre-compiled examples in doccano: https://bcgsc-doccano.herokuapp.com/
The following networks are available in NDEx (See link below)
Variant-Drug
Variant_Disease
Gene-Disease
Gene-Variant
http://public.ndexbio.org/#/user/77520e6b-816b-11e8-a4bf-0ac135e8bacf
The 9th international Semantic Web Applications and Tools for Life Sciences (SWAT4LS) conference is occurring concurrently with the CIViC Hackathon Dec 5-8, also at the NKI in Amsterdam. In an effort to cross-pollinate between our two meetings/hackathons efforts will be made to provide CIViC data for SWAT4LS hacking. Specifically we will attempt to incorporate CIViC evidence statements into WikiData to allow semantic web applications.
I'm interested in methods for testing the accuracy of information in variant databases, with an eye to developing ways to assess the accuracy and utility of AI technologies for variant interpretation. Hoping to find other folks interested in these issues as well, and to share ideas about how to approach this both conceptually and technically.
Ian King, University Health Network (Toronto)
Enter '0000000000' into the pubmed ID field and note that, even though the validation returned an error, the field does not show any error messages.
This topic will include how to create a bed-format file or VCF file of CIViC variants.
e.g. some alert for the database responsibles so they can dig into the reasons behind that disagreement and try to solve it
We currently have a small pencil next to the name to edit an evidence statement. One user stated that this wasn't intuitive, found it very frustrating. Maybe a bigger button, inside the evidence frame, that spells out edit this record
. This may be clearer than what appears to be a means of editing the label of the evidence item.
@ahwagner can you please edit the description with some more details?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
π Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. πππ
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google β€οΈ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.