SEA-PHAGES | Alphafold models of minor tail proteins

Link to this post \| posted 22 Oct, 2025 14:38
pchien	Hi Folks! We are busy annotated a few phage genomes and I've been following Krista Freeman's excellent powerpoint/information about using AF3 to model minor tail proteins. We are fairly confident we have found tail spike, Dit, and baseplate hub based on these models and Krista's guide. However, neither Dit nor baseplate hub seem to be listed on the official functions list. My major question is how we should proceed for calling them. I'd rather not just call them minor tail proteins when the structural prediction signature is super strong, but also appreciate that folks are working on a standardized approach. Thanks!

Link to this post | posted 22 Oct, 2025 14:38

Hi Folks! We are busy annotated a few phage genomes and I've been following Krista Freeman's excellent powerpoint/information about using AF3 to model minor tail proteins. We are fairly confident we have found tail spike, Dit, and baseplate hub based on these models and Krista's guide. However, neither Dit nor baseplate hub seem to be listed on the official functions list. My major question is how we should proceed for calling them. I'd rather not just call them minor tail proteins when the structural prediction signature is super strong, but also appreciate that folks are working on a standardized approach. Thanks!

Link to this post \| posted 22 Oct, 2025 14:51
debbie	Hi Peter, We are going to cautiously weigh in on what to name the various structures of siphoviruses. The data is coming to us fast and furious, and not all crystallographers are using the same nomenclature. I have added the features (as you have identified)to the Approved Function List that I think we can call with a modicum of confidence. Let me know if my list works! Thanks, debbie

Link to this post \| posted 22 Oct, 2025 14:55
pchien	This is great! The updated list is excellent. Do we want to include oligomer requirements like Krista had on there, like Dit should model as a hexamer, hub as a trimer, spike as a trimer?

Link to this post \| posted 22 Oct, 2025 15:00
debbie	Nope! But provide that here and I will add that to the function list, so others can pick the best model number. Would that work? Also provide me with the list of genes that you have called. I can add them as examples too. Thanks, debbie

Link to this post \| posted 22 Oct, 2025 15:05
pchien	will do (as soon as I finish with class today!)

Link to this post \| posted yesterday, 18:05
pchien	Sorry about the "best intention, worst execution" of not getting back to this thread Our class is annotating three new EG genomes and with a combination of Krista's presentation, Alphafold server, and use of Foldseek, we were able to confidently assign some functions. Specifically, we propose that BaconCheese40 is the distal tip complex (dit), BaconCheese42 is the baseplate hub protein, and BaconCheese43 is the tail spike protein. I made a series of slides to show this interpretation along with the assignments for orthologs from Mila11 and Slysloth, the other EG genomes we annotated. For dit, we went the extra step of confirming the tail width using the EM images of the phage to show that it was consistent with the outer diameter of the predicted dit hexamer complex. Hope this helps others! Edited yesterday, 18:08 2.71Mb

Link to this post | posted yesterday, 18:05

pchien

Sorry about the "best intention, worst execution" of not getting back to this thread smile

Our class is annotating three new EG genomes and with a combination of Krista's presentation, Alphafold server, and use of Foldseek, we were able to confidently assign some functions. Specifically, we propose that BaconCheese40 is the distal tip complex (dit), BaconCheese42 is the baseplate hub protein, and BaconCheese43 is the tail spike protein. I made a series of slides to show this interpretation along with the assignments for orthologs from Mila11 and Slysloth, the other EG genomes we annotated. For dit, we went the extra step of confirming the tail width using the EM images of the phage to show that it was consistent with the outer diameter of the predicted dit hexamer complex. Hope this helps others!

Edited yesterday, 18:08

Link to this post \| posted yesterday, 18:09
pchien	would love to know what you think!

Link to this post \| posted yesterday, 21:56
cdshaffer	Wow this is great work. Really good analysis on using alpha fold to discover the very likely localization and biological role for these proteins. I would say this goes way beyond what is expected for a typical annotation but really good work. Love the correlation of alphafold size predictions and physical observation in the EM. I only have one comment that really that has nothing to do with the role and function of these proteins but has more to do with nomenclature and a not that uncommon problem, so I think it is worth bringing up. If you look at the approved terms list you will see that Dit is mentioned twice. One approved term is just Dit and the other approved term is Dit/tail tip cage. Why is there two different terms and how would one know which one to pick? All this comes from the cryo EM work of Bxb1 done by Krista in the Hatfull lab. When she looked at gp32 in Bxb1 it is sitting right here where your protein is and she said that this Bxb1 protein is a fusion of two previously described proteins Dit and tail tip cage. Dit is the N terminal part of this Bxb1 protein and tail tip cage is the C term. So she named this with that double label. So I am seeing lots of good work that this protein is a Dit but I wonder, is it like Bxb1 and is it a fusion or is your phage the kind where the Dit and the tail tip cage are two separate polypeptides? I did not do an exhaustive search (sorry too much grading to do before grades are due), I did take Baconcheese 40 and do an HHPRED search to PDB, there was a very nice hit to the actual structure of that Bxb1 Dit/tail tip cage (9D93_Mf). Interestingly the alignment is not full length, only aa 75 to 275 of the 450 aa long BaconCheese protein aligned to aa 30 - 236 of the 685 aa Bxb1 Dit/tail tip cage. So the only part that is matching is the N term which is the Dit part. So my preliminary results would further support your annotation of just "Dit". If you want to dig deeper still, I would investigate what those unaligned amino acids at the C-term might be? They don't align to Bxb1 Dit/tail tip cage but they could align to some other tail tip cage or to something entirely different. Bottom line: I would certainly put my name on a paper that annotated this BaconCheese protein as Dit, but there might be something else going on here worthy of annotation at that C terminal end and a bit of further work might find it. On the other hand, an annotator always has to, at some point, say: "enough is enough, time to make the call" since there is not infinite time to investigate every protein. And so calling the "enough is enough" and annotating it Dit is totally fine in my book.

Link to this post | posted yesterday, 21:56

cdshaffer

Wow this is great work. Really good analysis on using alpha fold to discover the very likely localization and biological role for these proteins. I would say this goes way beyond what is expected for a typical annotation but really good work. Love the correlation of alphafold size predictions and physical observation in the EM.

I only have one comment that really that has nothing to do with the role and function of these proteins but has more to do with nomenclature and a not that uncommon problem, so I think it is worth bringing up.

If you look at the approved terms list you will see that Dit is mentioned twice. One approved term is just Dit and the other approved term is Dit/tail tip cage. Why is there two different terms and how would one know which one to pick?

All this comes from the cryo EM work of Bxb1 done by Krista in the Hatfull lab. When she looked at gp32 in Bxb1 it is sitting right here where your protein is and she said that this Bxb1 protein is a fusion of two previously described proteins Dit and tail tip cage. Dit is the N terminal part of this Bxb1 protein and tail tip cage is the C term. So she named this with that double label. So I am seeing lots of good work that this protein is a Dit but I wonder, is it like Bxb1 and is it a fusion or is your phage the kind where the Dit and the tail tip cage are two separate polypeptides?

I did not do an exhaustive search (sorry too much grading to do before grades are due), I did take Baconcheese 40 and do an HHPRED search to PDB, there was a very nice hit to the actual structure of that Bxb1 Dit/tail tip cage (9D93_Mf). Interestingly the alignment is not full length, only aa 75 to 275 of the 450 aa long BaconCheese protein aligned to aa 30 - 236 of the 685 aa Bxb1 Dit/tail tip cage. So the only part that is matching is the N term which is the Dit part. So my preliminary results would further support your annotation of just "Dit". If you want to dig deeper still, I would investigate what those unaligned amino acids at the C-term might be? They don't align to Bxb1 Dit/tail tip cage but they could align to some other tail tip cage or to something entirely different.

Bottom line: I would certainly put my name on a paper that annotated this BaconCheese protein as Dit, but there might be something else going on here worthy of annotation at that C terminal end and a bit of further work might find it. On the other hand, an annotator always has to, at some point, say: "enough is enough, time to make the call" since there is not infinite time to investigate every protein. And so calling the "enough is enough" and annotating it Dit is totally fine in my book.

Recent Activity

Alphafold models of minor tail proteins