I think I have found a pham that is annotated incorrectly for the functions, and I wanted to post what I found, just in case it may help someone else.

Pham 1929 (as of 9/4/17) in clusters L (Zakai_52) and O (Catdawg_81). In the cluster L, some of these are called membrane proteins or band-7 like, but I think they only have signal peptides, which can show up as transmembrane domains. Here is the TMHMM output for Zakai_52, but both Phobius and SignalP indicate signal peptides.

For comparison, Catdawg_81, in the same pham, has similar Phobius and SignalP results, but no called transmembrane domains by TMHMM. HHPred did have high probability and high coverage hits to Hfl proteins, which is probably why membrane protein was called:
High frequency of lysogenization C (HflC) family; SPFH (stomatin, prohibitin, flotillin, and HflK/C) superfamily
This model characterizes proteins similar to prokaryotic HflC (High frequency of lysogenization C). Although many members of the SPFH (or band 7) superfamily are lipid raft associated, prokaryote plasma membranes lack cholesterol and are unlikely to have lipid raft domains. Individual proteins of this SPFH domain superfamily may cluster to form membrane microdomains which may in turn recruit multiprotein complexes. Escherichia coli HflC is an integral membrane protein which may localize to the plasma membrane. HflC associates with another SPFH superfamily member (HflK) to form an HflKC complex. HflKC interacts with FtsH in a large complex termed the FtsH holo-enzyme. FtsH is an AAA ATP-dependent protease which exerts progressive proteolysis against membrane-embedded and soluble substrate proteins. HflKC can modulate the activity of FtsH. HflKC plays a role in the decision between lysogenic and lytic cycle growth during lambda phage infection.

Would it be helpful to have signal sequences noted in annotations, like we do for some domains?
Edited 05 Sep, 2017 04:35