Link to this post | posted 20 Dec, 2017 16:03
best

Link to this post | posted 20 Dec, 2017 14:13

best

I've been QC'ing annotations for three Cluster B2 genomes, and I decided to look a little deeper into the GTP Cyclohydrolase I protein (gp6 in all B2's to date). There are several annotations assigned to this gene in non-draft genomes: GTP cyclohydrolase GTP cyclohydrolase I FolE-like FolE Putative T-Fold protein The gene sits at the end of a run of genes associated with queuosine biosynthesis, so it is possible that it could be associated with queuosine. HHpred results indicate very good hits with GTP cyclohydrolase I (FolE) structures and to QueF structures. These are very similar proteins and folds, but they perform distinct transformations in the queuosine pathway, with QueF being specific to the pathway and GTP cyclohydrolase I producing a precursor to the pathway (i.e., the precursor can be shuttled to different pathways other than for queuosine). This paper is highly relevant to making the correct call: http://www.pnas.org/content/102/12/4264.full When looking at the alignments on HHpred, the motifs absent from QueF homologs and present in GTP cyclohydrolase homologs are present in the phage proteins. So that is a strike against the QueF-like designation. However, the fact that it is in an operon with QueCDE-like genes argues for a contextual basis for the QueF-like designation. See Figure 5 from the linked paper. The green QueF unique region is clearly like the cyclohydrolase group in the phage protein. The blue and yellow highlighted catalytic residues for HH and HxC are also in the phage protein. For now, the annotation should be "GTP cyclohydrolase I", which is a specific annotation, still associated with a precursor to the queuosine pathway, and avoids a simple gene name (FolE) derived from model bacteria. A "QueF-like" designation should not be added to the annotation, for instance, like the form of the annotations for the previous genes in the genome as found on the approved function list, "Queosine biosynthesis protein, QueE-like". This would be tempting given the hit probabilities in HHpred, but it would be inaccurate. Finally, it could be interesting to clone this gene at some point and do some old fashioned biochemistry. In this more recent paper, http://www.pnas.org/content/113/11/E1452.full, it is noted that queF-like genes have been identified in other phage. Novel cross reactivity? Who knows. The function "GTP cyclohydrolase I" should be added to the approved function list. Edited 20 Dec, 2017 14:14

Link to this post | posted 11 Feb, 2017 18:23

best

Hi Chris, I was able to easily install the byphagewithbase branch on the 2017 SEA VM, and it runs. However, the output produced is not identical to your pre-run reports with respect to the text. On my installation, I get "Phages represented in each track" output, but that is where the report ends. I don't get "Summary of Final Annotations" sections. Any thoughts? In the instructions above, there was no need to install PyPDF2 or requests. They were already there in the default configuration. Also, one other note on the output. If you could add the track number to the tracks, that would make it easier to identify them rather than having to count down the page. Definitely don't lose the phage names on the track, but maybe begin each of those lines with "track_number: phage name + X".

Link to this post | posted 11 Feb, 2017 17:58
best	Thanks, Chris. I'll give it a go with your latest code and let you know if I run into any errors.

Link to this post | posted 10 Feb, 2017 14:07

best

Hi Chris, Thanks for working on the starterator program and updating the output. It looks great. The idea of pre-running on the phams and posting that output is excellent. Very helpful and likely how we will use the functionality most often. Would you be able to add a post (or add to github) the command line procedure for updating to the current version of starterator in the 2017 SEA VM? Thanks, Aaron