Annotation source: Pfam from a HMMER search (no GO/InterPro) for: Protomyces inouyei (Protinou). Other annotated species: UniProt (GO, Pfam, InterPro). For the species above, an empty GO or InterPro field means 'not searched', not 'no annotation'.
Run summary
Counts cover every protein in the presence matrix, independent of the filters below.
Novelty candidates
Proteins in matrix
Ingroup
Outgroup
Functionally annotated
Novelty candidates per ingroup species
Proteins originating in each ingroup species that pass the novelty filter.
How to read this page
- Candidate. A protein from an ingroup proteome that has a qualifying hit in at least the stated fraction of the ingroup proteomes (its own proteome counts) and no qualifying hit in any outgroup proteome. Counts are proteins, not gene families.
- Qualifying hit. A protein-search hit that passes the E-value cutoff (default 1e-5), is not explained away by the query's own in-genome paralog (a hit is dropped when the paralog hits the same target better, unless the hit is itself stronger than 1e-20), and passes the optional query-coverage floor when one was set. "No qualifying hit" is not proof of absence: a filtered or weak hit is not counted.
- Gene family. Candidates are clustered with mmseqs before TBLASTN. A family groups candidates from different ingroup species that cluster together. A single protein is its own family.
- TBLASTN column. The number of outgroup genomes where TBLASTN of the cluster representative found a hit at the E-value cutoff. The result is copied to every cluster member. It is evidence only: it does not remove a candidate. A TBLASTN hit with no protein hit often means an unannotated or mis-annotated outgroup gene.
- Outgroup signal (when the run recorded it). The strongest outgroup evidence by query coverage, including hits that were filtered out: none (no hit), domain_only (best coverage below the threshold, often one shared domain) or broad (at or above it). The threshold box on this page changes the split. The default of 50% is the median seen in one run, not a validated cutoff. Protein coverage is a single alignment; TBLASTN coverage is the union of all aligned spans in a genome, so the two differ in kind. Evidence only: it never changes the candidate list.
- What to check next. Rows with a TBLASTN hit or a broad outgroup signal may be old families with a lineage-specific member or annotation gaps. Open the alignment (TBLASTN cell) and the Outgroup signal detail before treating a row as novel.
Present (protein search)
TBLASTN hit (outgroup genome)
No qualifying hit