Feature hasn't been suggested before.
Describe the enhancement you want to request
Would you be interested in merging a database skill for individual-level queries
over the 1000 Genomes Project cohort (3,202 WGS samples, GRCh38) ?
It answers: which individuals carry variants matching given criteria in a region
with certain zygosity; which variants a set of individuals carries; which
individuals are homozygous-reference at a position; kinship between two
individuals; cohort population and pedigree metadata. Variants are filterable by
gnomAD v4.1 AF, AlphaMissense score/class, ClinVar significance, VEP consequence.
For example:
Which individuals in the cohort carry charge-reversal missense variants (e.g. Asp→Lys) at
the MCM2/MCM5 interface of the replicative helicase, and do any of them also carry a
reciprocal charge-reversal on the opposing subunit that would restore the electrostatic
pairing?
The agent answers this by selecting carriers at each subunit interface and intersecting the
sample sets — no other database required.
More examples where it's essential: https://github.com/dnaerys/onekgpd-mcp/tree/master/examples
The skill queries a public endpoint we maintain (https://dnaerys.org/online/) - open, no key,
no rate limit. We tested it in an OpenScience instance and confirmed it works as expected.
Implemented in the standard agent-skills format: https://github.com/dnaerys/onekgpd-skill
Feature hasn't been suggested before.
Describe the enhancement you want to request
Would you be interested in merging a database skill for individual-level queries
over the 1000 Genomes Project cohort (3,202 WGS samples, GRCh38) ?
It answers: which individuals carry variants matching given criteria in a region
with certain zygosity; which variants a set of individuals carries; which
individuals are homozygous-reference at a position; kinship between two
individuals; cohort population and pedigree metadata. Variants are filterable by
gnomAD v4.1 AF, AlphaMissense score/class, ClinVar significance, VEP consequence.
For example:
The agent answers this by selecting carriers at each subunit interface and intersecting the
sample sets — no other database required.
More examples where it's essential: https://github.com/dnaerys/onekgpd-mcp/tree/master/examples
The skill queries a public endpoint we maintain (https://dnaerys.org/online/) - open, no key,
no rate limit. We tested it in an OpenScience instance and confirmed it works as expected.
Implemented in the standard agent-skills format: https://github.com/dnaerys/onekgpd-skill