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

DNA-binding ferritin-like protein

| posted 20 May, 2017 16:29
The function "DNA-binding ferritin-like protein" is in Streptomyces phage Jay2Jay (BE1), but isn't on the approved list. Should we keep this in the other BE1 with the same gene call or have it removed from Jay2Jay?

| posted 04 Aug, 2022 01:36
On a similar topic -

I am reviewing PumpkinSpice_33 in PECAAN (stop site 15,395).


Several genes in the same pham, including cluster member MindFlayer_33, call DNA binding protein as the function (though Jay2Jay, mentioned in the previous post, ended up as NKF).

It seems like there is some CDD evidence to support this. There is one CDD hit which mentions nonspecific DNA binding: cd01043 "DPS protein, ferritin-like diiron-binding domain."
CDD hit stats: 32% identity 44% coverage 93% probability

And another: COG0783 "DNA-binding ferritin-like protein"
CDD hit stats: 25% identity. 41% coverage 93% probability

Are these sufficient evidence to call DNA binding protein?

Edited 04 Aug, 2022 01:37
| posted 04 Aug, 2022 20:43
We worked on many of the proteins in this pham. I don't think any student has worked on this protein in an in-depth investigation so I spent about 30 minutes poking at the literature and looking at a member of this pham we have been working on this summer pepperwood_draft_32 (stop: 14017). Looking broadly at the hhpred and blast hits, there is clearly a very wide ranging family of proteins all like those in this pham. Also, pepperwood_draft_32 may not be the best member to represent the group as it is a bit shorter than the others ( but its annotated with the longest ORF).

In looking over the paper associated with the high matching crystal 2C2U, I can see that many of the proteins in this family have dual functions, DNA binding and Iron sequestration. Most of these proteins appear to be expressed in bacteria in responce to starvation or other forms of stress and are probably play a role in protecting DNA from damage during high levels of oxidative stress. The general name for proteins like this are DPS for DNA Protection during Starvation.

In the paper on the 2C2U crystal they stated that the mechanism for DNA binding was not well understood, so i looked for recent papers and found this one:

it appears to do a good job of summarizing these DPS proteins if you really want to get into the literature. But the one take home message I got from this is that proteins in this wide family have a variety of activities some do and some don't have DNA binding, some do and some don't store Iron, and some do and some don't have ferroxidase activity. So bottom line is that it is really going to be hard to get a good convincing argument of what function any particular protein has.

So Bottom line I would say is that annotation in this case is about which kind of error you think is worse, false positives or false negatives. For functional annotation I prefer false negatives since too many people will take any annotation as "the truth". Many naive users may not understand that an annotation might mean something that is "weak but interesting". This propagation of incorrect functional annotations is a well understood problem with large databases like genbank and I would rather say nothing for the function and not exacerbate the problem. However, as you cited above, there is clearly some evidence to suggest a DNA binding activity for this protein and so it is totally reasonable to come to the conclusion that there is enough evidence to support this annotation. What we really have here is a situation where reasonable and well trained annotators can disagree and that is OK, without clear wet bench experimentation annotations are always just well informed estimations and it is totally reasonable to have different estimations.
| posted 04 Aug, 2022 21:28
Chris, thank you so much for the info and your advice on false negatives. I decided to leave it as NKF based on that rationale. It's helpful to hear about your investigation process too. Thank you!
| posted 05 Aug, 2022 05:36
I agree with Chris that it would be premature to call this gene product a ferritin-like DNA binding protein, a ferritin, or perhaps even a DNA binding protein. However, I would argue that because it is readily identifiable a member of the well-established DPS protein family, it would be beneficial to the community if it were to be annotated as a "DPS family protein." The advantage of using this designation over something like "hypothetical protein" or even "DNA binding protein" is that it would draw attention to the fact that it's a member of an established protein family, and that there are therefore reasonable hypotheses for future researchers about its potential range of functions. We frequently identify proteins as members of families without specifying their functions, and this example strikes me as no different.

| posted 09 Aug, 2022 14:47
I guess I am going to weigh in as grateful for this post. We have a record and that will help as we go forward.
For now, I am going to suggest that we do not add DPS family protein into the approved list.
I would recommend that this protein be called a Hypothetical Protein for now.
Login to post a reply.