Link to this post | posted 05 Jun, 2019 18:04

cdshaffer

In situations like this we are left with the very hand wavy answer that it is simply that the functions provided by the holins are being provided by proteins which cannot be detected with bioinformatic based tests. This can happen by either (1) convergent evolution (i.e. in this context, two completely different proteins evolving to carry out the function) or that (2) a completely different biochemical pathway has evolved in the B1 phages to "solve" the problem of how to lyse the host cell wall In either situation bioinformatics fails us (by definition really) and we are left going back to the wet bench to answer the question "how do B1 phage lyse their host?" So maybe someone else might be aware of published or unpublished results from bench work in this area and might have further insights.

Link to this post | posted 28 May, 2019 20:12

cdshaffer

I have just returned from vacation and have updated the the starterator pages to the most recent version of the phamerator database which was posted this morning. I checked bunnies gene 32, it is now in pham 39047 and appears to be in sync across pecaan, starterator, phagesdb.org and phamerator.org. Checking phage Fulbright shows some discrepancies across the various platforms but that could be syncronization issues since the last update was posted so recently. I would give it a day and if reports are still missing please repost and I can look into it.

Link to this post | posted 10 May, 2019 21:57

cdshaffer

Just a heads up that I will be out of the country for the next two weeks and will have unknown quality of internet connections throughout my travels. As such, it is likely I will not be able to update online starterator reports in a quick and timely fashion. Most pham groups remain static and the old report should suffice in most cases. However, should an updated version of the pham database move your gene of interest to a new pham number the virtual machine remains a solution if you are in a hurry. See this page for help. Thank you for your patience and understanding. Edited 10 May, 2019 21:59

Link to this post | posted 10 May, 2019 21:44
cdshaffer	there is a button that should do that but if the button on each gene is not working the whole phage might not either. Go to Admin menu -> phages; search for your phage and click the edit button. From there you should get a page which includes button to reblast the whole phage.

Link to this post | posted 06 May, 2019 20:10

cdshaffer

We have found a few enzymes in some of our streptomyces phages where the "like" part of the approved function list did not apply. For these we posted a request for a "new" function without the "like" part. In support we provided references to the papers describing the novel structure of the protein as evidence that our hiqh quality HHPRED hits to these crystals mean the more generic term should be added to the list. However in this case for Jacko 31 I think the "NrdH-like glutaredoxin" is appropriate since on PECAAN results there is an HHPRED hit with >98% prob and >98% coverage to 4FIW_A which is the published crystal structure of NrdH from E coli. Given this hiqh quality hit by HHPRED to NrdH I would say that the term "NrdH-like" does apply.

Link to this post | posted 18 Apr, 2019 15:05
cdshaffer	I just tried online phamerator.org and was successful so it might have been a temporary issue. If it is still failing for the students I would suggest trying to create a new account as a work around, also posting to the chat on the phamerator page can't hurt.

Link to this post | posted 10 Apr, 2019 17:30

cdshaffer

Lee is correct about the presence of non-printable or control characters in your notes field, if you see this line in your documentation: /note="" then you know there is something in the notes which will mean "hypothetical protein" will not be added correctly. I see this a lot and I always assumed it was an instance of a very common problem in bioinformatics that has to do with newline definitions of mac vs. dos vs. windows vs. linux (its complicated but see this wiki page if you are curious). If my suspicion is correct then PECAAN is not the problem and cannot fix anything, but rather its an issue with one of the various specific programs you use and/or your OS as they handle text as you process the downloaded file. The solution I have found is to use the "recreate" button in DNA Master and then parse again. This has fixed the issue for me. So the protocol then is: 1. Paste in the downloaded file into the documentation window 2. Parse this entry by clicking the "parse" button, the next "parse" button, then "yes" 3. Recreate the documentation by clicking the "recreate" button, then "yes" 4. Now parse this version of the documentation by following step 2 again I usually do one more round of recreate and parse just to be sure. So the click stream I use is: paste, parse, parse, yes, recreate, yes, parse, parse, yes, recreate, yes, parse, parse, yes Edited 10 Apr, 2019 17:49

Link to this post | posted 21 Mar, 2019 15:47

cdshaffer

To me annotations like helix-turn-helix DNA binding domain should be added only if there is sufficient evidence that the protein really does have a domain of that type AND, more importantly, there is not a more specific approved term that is also supported by the evidence. For example many (if not all??) sigma factors contain HTH domains but "sigma factor" so either term would apply. However sigma factor is a much better annotation than HTH binding domain protein since it is a more specific term. So you have to look at the evidence with an eye toward the validity of "sigma factor" vs "HTH domain", this is why each match should be evaluated with respect to the size and location of the exact match with respect to the whole proteins. Full length alignments are much better than short little domain matches, but if all you have is a short high quality match to an HTH domain then I would add it. I think for short domain matches I would also focus on the HHPRED results. A "helix-turn-helix domain" is really annotating the presence of a structural domain so I would want to focus on the programs that are trying to find similarity at the structural level (HHPRED, Phyre2 etc), not the primary amino acid level (i.e. BLASTP) In this particular case since FIC has been rejected as a term, there is no better approved term that HTH domain. Thus, I would just evaluate the evidence from hhpred as to the question do I really have a HTH domain and add it if I felt the evidence justified it.

Link to this post | posted 20 Mar, 2019 16:23

cdshaffer

Hmmm, that pham no longer exists, probably was in the old database but not the newest one, if you are using the virtual machine it is important to run phamerator first to check for and install any database updates before running starterator. I have used the most recent "stable" version of starterator which has a few udpates that you have not seen if you are used to the whole phage report from the version in the virtual machine, so just be aware that the output will be a bit different. I have posted the results here: https://wustl.box.com/s/irzr8fel3z6e1tmr9qirymj3fl91j0ng You should also know that most users no longer bother with the whole phage reports but instead go to the pre-computed online versions of the per pham starterator reports. You can see in this article how to get access to those online reports. I always try to keep these reports up to date with new versions of the database.

Link to this post | posted 12 Mar, 2019 18:02

cdshaffer

I think it is always better to be as specific as the evidence allows. The three terms you cite are not inconsistent just different levels of specificity. I think most, if not all, kinases use ATP as the source of the phosphate so not surprising that a kinase has an identifiable "ATP binding cassette". I would pay particular attention to the length of the alignments, does a particular alignment include the majority of your protein? is that alignment along the majority of the subject? Thoese "full length by full length" alignments are the most informative and I would pick the most specific term (i.e. the thymidylate kinase). On the other hand if the region of your protein that matches the "ATP binding cassette" is the same region that is matching to a subpart of some thymidylate kinase then you likely just have an ATP binding domain that is particularly similar to the ATP binding domain found in a thymidylate kinase, in that case I would use the more general ATP binding domain.