Pasteur studying a flask, waiting to see what grows. Research on unresolved cases needs the same patience, and the discipline not to call it early.
A lead is not a finding.A rank is not a probability.
Tulparus keeps candidate genes on their own list, re-checks open cases when knowledge changes, and records the versions behind every result, so a research question stays a question until the evidence closes it.
New genes on their own list.Never mixed with the known.
Genes without an established disease link are ranked separately, with the reasons each one is there, so a research lead is never dressed up as a diagnosis.
Every release, a second look.Every change, on record.
When a knowledge releaseA dated snapshot of the reference databases and literature the models read from, so every result can name what it was based on. changes what is known about a gene, the open cases that depend on it are re-checked, and what moved is shown with the release that moved it.
Same inputs, same answer.Six months later.
Every result records the case, model, knowledge and pipeline versions behind it. An analysis can be repeated as it was, or rerun under newer versions with the difference shown.
We publish what we measure.And where it stops.
The models are evaluated on named sets for each release, and the figures come with their sample size and interval. The research notes say how each was tested and what it cannot do.