The official website of the HHMI Science Education Alliance-Phage Hunters Advancing Genomics and Evolutionary Science program.

Welcome to the forums at Please feel free to ask any questions related to the SEA-PHAGES program. Any logged-in user may post new topics and reply to existing topics. If you'd like to see a new forum created, please contact us using our form or email us at

Membrane protein

| posted 18 Jan, 2019 16:47
Does a gene that BLASTs to Membrane protein, Band-7-like need to have multiple transmembrane domains to call it Membrane protein, Band-7-like?
Edited 18 Jan, 2019 19:35
| posted 19 Jan, 2019 13:55
No. Band-7 proteins are their own class.

Thanks for asking!
| posted 30 Jan, 2019 21:13
Would it be possible to get greater clarification for the function assignment 'membrane protein?'

The online bioinformatics guide states the following "a protein can be assigned the function "membrane protein" based on the results of the programs TMHMM and/or SOSUI. We use this assignment if we find two or more potential membrane domains predicted by one of the programs, or, a single membrane domain predicted with confidence by both programs."

The official function list seems to require 2 confirmed transmembrane domains for the function 'membrane protein' to be called.

Is it acceptable to assign the function membrane protein with only 1 TM confirmed by TMHMM and SOSUI?
Thank you for any guidance!

| posted 01 Feb, 2019 03:50
Hi Denise,
If you find one transmembrane domain with TMHMM, you can only call it a membrane protein if you can validate it with a different transmembrane finder (such as SOSUI). (If two different algorithms detect a trans membrane domain, you can hopefully rule out false positives.)
Does that clarify it?
| posted 01 Feb, 2019 19:23
Yes - I just hesitated when I saw the new note associated with the function 'membrane protein' in the Official Function List at It reads "at least two (2) transmembrane domains found using TMHMM." My group was assigning the function 'membrane protein' using the 'rules' outlined in the Bioinformatics Guide and which you re-iterated above. The clarification is helpful. Thank you!
| posted 11 Mar, 2019 18:36

SOSUI is down. We contacted the host, and we do not know when it will be back up. We see that TMHMM predicts one transmembrane domain. The Official Function list says we need two domains called by TMHMM. This means we cannot call it a membrane protein, correct? Why does it have to have two domains?
| posted 15 Mar, 2019 02:54
I don't think it has to have 2 domains, it just has to have 1 convincing one. The problem is that one program doesn't have adequate sensitivity. So if there is only one, we want it verified by 2 sources. Without the option to 'verify' a transmembrane domain with a second program, let's not add it as a function.
| posted 28 Mar, 2019 20:41
How long does a membrane protein need to be? I have run into a couple of cases where the protein is very small (30-45ish aa) and has one transmembrane domain, called by both TmHmm and SOSUI. Or are these better left as NKF? Here's an example:

From TmHmm:
# WEBSEQUENCE Length: 43
# WEBSEQUENCE Number of predicted TMHs: 1
# WEBSEQUENCE Exp number of AAs in TMHs: 19.16541
# WEBSEQUENCE Exp number, first 60 AAs: 19.16541
# WEBSEQUENCE Total prob of N-in: 0.01540
# WEBSEQUENCE POSSIBLE N-term signal sequence
WEBSEQUENCE TMHMM2.0 outside 1 4
WEBSEQUENCE TMHMM2.0 inside 25 43

This amino acid sequence is of a MEMBRANE PROTEIN which have 1 transmembrane helix.
No. N terminal transmembrane region C terminal type length
| posted 28 Mar, 2019 23:51
I do't believe a sequence has to be very long to contain a transmembrane domain. for me the 2 sides of the coin are: 1. is one membrane real? (so we ask that you can find it with 2 different finder programs) 2. It appears that defense systems (toxins, for example) can be 'attached' to the cell membrane, so that membrane protein might be a good place to look into that if you have a phenotype that suggests that phage has provided a defense. i.e. it would just be good to know. It will also be found out once you have a phenotype to investigate, so there is no harm by not calling it.
Not helpful, right? So, I think my take on this is if you take the time look and confirm, report it. If this is all you miss in the genome, it is small potatoes.
| posted 29 Mar, 2019 13:10
Hi Debbie,
This actually was quite helpful, and for now I will be probably be assigning these as NKF. But it may form the basis for an senior thesis project I will be co-advising next year with Mike Buckholt.
Login to post a reply.