Chapter three
The map of regimes
Die Behauptung
The nine cells are not nine separate scores, they are a map of which evidence wins where. The rows are how much we already know about a protein (nothing, a little, or a fair amount); the columns are the three branches of the Gene Ontology (molecular function, biological process, cellular component). Read across the map and a pattern appears: a learned classifier carries the proteins we know nothing about; what we already know about a protein, its associations and its lineage in the ontology, carries the partly-known ones; and plain sequence similarity proposes candidates but cannot tell the good from the bad on its own. First in seven of the nine cells is the shape of that map, not one number, and the two cells we do not hold are exactly the two the map predicts are hardest.
Ziehen Sie die Markierung, um den Beleg und die Operation zu sehen, die die Zahl neu erzeugt.
Die Belege
| Regime | ||
|---|---|---|
| NK molecular function | carried | |
| NK biological process | carried | |
| NK cellular component | carried | |
| LK molecular function | carried | |
| LK biological process | frontier | not won |
| LK cellular component | carried | |
| PK molecular function | carried | |
| PK biological process | frontier | not won |
| PK cellular component | carried |
Unsere Vorbehalte
Wir nennen sie, bevor sie jemand für uns nennt.
- The regime reading, which evidence carries which cell, comes from a leave-one-family-out ablation whose receipt lives in the lab repository rather than in this page storage snapshot. The values themselves are withdrawn while the campaign recomputes, so what this table publishes today is the reading and not the measurement.
- That homology proposes but does not discriminate is a statement about this reference and this frame. A different reference proteome could shift it.