SEA-PHAGES | All posts created by cdshaffer

← previous
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
next →

Link to this post \| posted 12 Mar, 2021 18:47
cdshaffer	Your question about false positives is an interesting one. I had always assumed the algorithms were specifically designed to distinguish the differences between membrane domains and simple hydrophobic helices. So I went back to the 2001 paper for TMHMM (doi:10.1006/jmbi.2000.4315). The intro in the paper has a really good discussion on the early methods used to distinguish just that issue. There is also a whole section of the paper on this issue. Bottom line is there are other structures and length requirements that help in the determination which helps distinguish a "real" transmembrane domain. Might even be worth pointing out this paper to students who are interested, if only to read the intro. As for the issue of false positives with TM-HMM, according to the paper, the algorithm has a specificity of around 99% if there is not a leader peptide, so I think the protocol as defined is a pretty good one and further support from BLAST is not required. But this issue that a leader peptide reduces the quality of the results is very interesting. Maybe the SEA-phages protocol should be amended if a leader peptide is predicted. That is really a good question for a faculty workshop I think. Edited 12 Mar, 2021 18:52

Link to this post | posted 12 Mar, 2021 18:47

Your question about false positives is an interesting one. I had always assumed the algorithms were specifically designed to distinguish the differences between membrane domains and simple hydrophobic helices. So I went back to the 2001 paper for TMHMM (doi:10.1006/jmbi.2000.4315). The intro in the paper has a really good discussion on the early methods used to distinguish just that issue. There is also a whole section of the paper on this issue. Bottom line is there are other structures and length requirements that help in the determination which helps distinguish a "real" transmembrane domain. Might even be worth pointing out this paper to students who are interested, if only to read the intro.

As for the issue of false positives with TM-HMM, according to the paper, the algorithm has a specificity of around 99% if there is not a leader peptide, so I think the protocol as defined is a pretty good one and further support from BLAST is not required.

But this issue that a leader peptide reduces the quality of the results is very interesting. Maybe the SEA-phages protocol should be amended if a leader peptide is predicted. That is really a good question for a faculty workshop I think.

Edited 12 Mar, 2021 18:52

Posted in: Annotation → Membrane proteins

Link to this post \| posted 11 Mar, 2021 17:39
cdshaffer	Yes I agree, I was not trying to imply otherwise. My thought experiment was more with the idea of a outside reviewer. There are many papers out there that talk about how poor the annotations are in genbank as a whole, so I could imagine a naive reviewer liking the "published set" over and above the "all genbank set". This brings up the point that one might pick all SEA-PHAGES as a test set over all phages in genbank or all published phage if annotation consistency was important to the analysis. Again, the experimenter should pick the best possible dataset for the question.

Link to this post | posted 11 Mar, 2021 17:39

cdshaffer

Yes I agree, I was not trying to imply otherwise. My thought experiment was more with the idea of a outside reviewer. There are many papers out there that talk about how poor the annotations are in genbank as a whole, so I could imagine a naive reviewer liking the "published set" over and above the "all genbank set".

This brings up the point that one might pick all SEA-PHAGES as a test set over all phages in genbank or all published phage if annotation consistency was important to the analysis. Again, the experimenter should pick the best possible dataset for the question.

Posted in: General Message Board → Comparative analysis

Link to this post \| posted 10 Mar, 2021 18:03
cdshaffer	to me this is mostly an issue of good experimental design and picking the right dataset for your experiments. So consider these two possible sentences you could write in a hypothetical paper and decide which one would be better at convincing the reviewer to accept your paper and its conclusions: 1. "To cast a wide net and compare as many as possible we collected and analyzed all phage in genbank" 2. "To ensure the highest quality of data and gene annotations we selected only phage genomes which have been published in peer-reviewed journals" if I were a reviewer I would be fine with option 1 if the study was mostly about sequence variation with little or no input from annotations. The more your experimental conclusions rely on annotations the more I would favor option 2. As a middle ground, you could also consider all the phage in the refseq database instead of all of genbank.

Link to this post | posted 10 Mar, 2021 18:03

cdshaffer

to me this is mostly an issue of good experimental design and picking the right dataset for your experiments. So consider these two possible sentences you could write in a hypothetical paper and decide which one would be better at convincing the reviewer to accept your paper and its conclusions:

1. "To cast a wide net and compare as many as possible we collected and analyzed all phage in genbank"

2. "To ensure the highest quality of data and gene annotations we selected only phage genomes which have been published in peer-reviewed journals"

if I were a reviewer I would be fine with option 1 if the study was mostly about sequence variation with little or no input from annotations. The more your experimental conclusions rely on annotations the more I would favor option 2. As a middle ground, you could also consider all the phage in the refseq database instead of all of genbank.

Posted in: General Message Board → Comparative analysis

Link to this post \| posted 06 Mar, 2021 21:29
cdshaffer	I would always prefer the HHPRED matches (if I find them) over the blast results. This is due, in no small part, on the quality of the different databases being searched and the relative sensitivity of the algorithms. The source for many of these "discrepancies" like your list is that the alignments are only matching to part of your protein or to just part of the subject. Since some proteins have multiple functional parts all connected together in a single polypeptide chain this can lead to what I would call a "partial annotation". Also note that your first two possibilities do not really "disagree" they really are just different levels of specificity. When trying to decide on levels of specificity, I first direct my students to try to understand the differences in what the terms mean, good sources include the sea-phages approved terms list, the EXPASY enzyme class list, Wikipedia, intro bio text books etc. are all good sources. Once you have a better understanding of the terms you can then look for evidence to help you decide if a higher level of specificity is justified or not. As for this particular protein, if I scan through the top 15-20 hits from prokaryotes (i.e. I am going to ignore the two human mitochondrial proteins) I see many hits to proteins that are described to have BOTH a helicase activity AND a nuclease activity. This explains the "discrepant" results, so the question becomes: does this protein from crewmate also have those two domains or just one. This is why you see annotators often talk about the size of the protein and the size of the match. I quickly focus on the length of the alignment and which part of the subject is matching. Most of these alignments cover about 75% of crewmate 28 but you can see that they only match a much shorter part of the subjects that are described to have both a nuclease and a helicase activity (like residues 1005-1232, 790-1014 or 129-368 ). So likely crewmate is similar to either the nuclease or the helicase part of these larger proteins, but I cannot tell which based simply on the summary data presented in the table. Looking at the other matches I can see hits to pfam domains and "cd conserved domains" that are all different types of exonucleases. So what we have here is likely an exonuclease that is often found as part of a large helix/nuclease combo protein. So I am pretty convinced that either of the first two options on your list could be appropriate here. When talking about this with my students I would point out that since they are the first author it is really up to them to read up on the two terms and decide if they think the "cas4" is better than the generic "exo" but I would be willing to put my name as an author on either of those annotations since one is just a more specific subtype of the other.

Link to this post | posted 06 Mar, 2021 21:29

cdshaffer

I would always prefer the HHPRED matches (if I find them) over the blast results. This is due, in no small part, on the quality of the different databases being searched and the relative sensitivity of the algorithms. The source for many of these "discrepancies" like your list is that the alignments are only matching to part of your protein or to just part of the subject. Since some proteins have multiple functional parts all connected together in a single polypeptide chain this can lead to what I would call a "partial annotation".

Also note that your first two possibilities do not really "disagree" they really are just different levels of specificity. When trying to decide on levels of specificity, I first direct my students to try to understand the differences in what the terms mean, good sources include the sea-phages approved terms list, the EXPASY enzyme class list, Wikipedia, intro bio text books etc. are all good sources. Once you have a better understanding of the terms you can then look for evidence to help you decide if a higher level of specificity is justified or not.

As for this particular protein, if I scan through the top 15-20 hits from prokaryotes (i.e. I am going to ignore the two human mitochondrial proteins) I see many hits to proteins that are described to have BOTH a helicase activity AND a nuclease activity. This explains the "discrepant" results, so the question becomes: does this protein from crewmate also have those two domains or just one. This is why you see annotators often talk about the size of the protein and the size of the match. I quickly focus on the length of the alignment and which part of the subject is matching. Most of these alignments cover about 75% of crewmate 28 but you can see that they only match a much shorter part of the subjects that are described to have both a nuclease and a helicase activity (like residues 1005-1232, 790-1014 or 129-368 ). So likely crewmate is similar to either the nuclease or the helicase part of these larger proteins, but I cannot tell which based simply on the summary data presented in the table. Looking at the other matches I can see hits to pfam domains and "cd conserved domains" that are all different types of exonucleases. So what we have here is likely an exonuclease that is often found as part of a large helix/nuclease combo protein. So I am pretty convinced that either of the first two options on your list could be appropriate here.

When talking about this with my students I would point out that since they are the first author it is really up to them to read up on the two terms and decide if they think the "cas4" is better than the generic "exo" but I would be willing to put my name as an author on either of those annotations since one is just a more specific subtype of the other.

Posted in: Functional Annotation → Phage gene annotation has matching phage genes have 4 different proteins - which one is a match?

Link to this post \| posted 27 Feb, 2021 19:03
cdshaffer	in the database there are 11 phage genes with subcluster B1 they all appear in current pham 48591: `\| phageid \| subcluster \| gene name \| notes \| phamid \| \| BatteryCK \| B1 \| 14 \| holin \| 48591 \| \| Beaglebox \| B1 \| 14 \| holin \| 48591 \| \| DoesntMatter \| B1 \| 15 \| holin \| 48591 \| \| LeeLot \| B1 \| 15 \| holin \| 48591 \| \| Magic8 \| B1 \| 15 \| holin \| 48591 \| \| Megatron \| B1 \| 15 \| holin \| 48591 \| \| Mosaic \| B1 \| 14 \| holin \| 48591 \| \| OliverWalter \| B1 \| 15 \| holin \| 48591 \| \| ProfessorX \| B1 \| 15 \| holin \| 48591 \| \| Trypo \| B1 \| 15 \| holin \| 48591 \| \| Xavier \| B1 \| 14 \| holin \| 48591 \|` for the totally geeky here is the mysql query I used: `> select gene.phageid, phage.subcluster, gene.name as "gene name", gene.notes, gene.phamid from gene join phage on gene.phageid=phage.phageid where phage.subcluster="B1" and gene.notes like "%holin%";` Edited 27 Feb, 2021 19:04

Link to this post | posted 27 Feb, 2021 19:03

cdshaffer

in the database there are 11 phage genes with subcluster B1 they all appear in current pham 48591:


| phageid      | subcluster | gene name | notes | phamid |

| BatteryCK    | B1         | 14        | holin |  48591 |
| Beaglebox    | B1         | 14        | holin |  48591 |
| DoesntMatter | B1         | 15        | holin |  48591 |
| LeeLot       | B1         | 15        | holin |  48591 |
| Magic8       | B1         | 15        | holin |  48591 |
| Megatron     | B1         | 15        | holin |  48591 |
| Mosaic       | B1         | 14        | holin |  48591 |
| OliverWalter | B1         | 15        | holin |  48591 |
| ProfessorX   | B1         | 15        | holin |  48591 |
| Trypo        | B1         | 15        | holin |  48591 |
| Xavier       | B1         | 14        | holin |  48591 |

for the totally geeky here is the mysql query I used:

> select gene.phageid, phage.subcluster, gene.name as "gene name",
gene.notes, gene.phamid from gene join phage on gene.phageid=phage.phageid
where phage.subcluster="B1"
and gene.notes like "%holin%";

Edited 27 Feb, 2021 19:04

Posted in: Cluster B Annotation Tips → Holin

Link to this post \| posted 19 Feb, 2021 18:12
cdshaffer	The current versions of both chrome and firefox have stopped support for FTP. I started up an old version of the SEA-VM and used the really old version of firefox therein to confirm that the FTP server is up and running. So to access it directly you will need to use an FTP client to connect. In the past, on Windows machines, I have used Filezilla and cyberduck to access FTP sites. If you have a mac, you can also just connect using the finder -> go menu -> connect to server -> enter address of ftp://cobamide2.bio.pitt.edu/ -> click connect -> select connect as guest => click connect. This will open a finder window which is connected to the ftp server. Be patient with file loading it can be slow; the file you want (dna master.exe) is in the DNAMas folder. You can use typical drag and drop to copy the file. Final step, of course, is to transfer the file to your windows machine by whatever method you find most convenient.

Link to this post | posted 19 Feb, 2021 18:12

cdshaffer

The current versions of both chrome and firefox have stopped support for FTP. I started up an old version of the SEA-VM and used the really old version of firefox therein to confirm that the FTP server is up and running. So to access it directly you will need to use an FTP client to connect. In the past, on Windows machines, I have used Filezilla and cyberduck to access FTP sites.

If you have a mac, you can also just connect using the finder -> go menu -> connect to server -> enter address of ftp://cobamide2.bio.pitt.edu/ -> click connect -> select connect as guest => click connect. This will open a finder window which is connected to the ftp server. Be patient with file loading it can be slow; the file you want (dna master.exe) is in the DNAMas folder. You can use typical drag and drop to copy the file.
Final step, of course, is to transfer the file to your windows machine by whatever method you find most convenient.

Posted in: DNA Master → DNA master server down?

Link to this post \| posted 17 Feb, 2021 16:53
cdshaffer	The issues you are having are related to the restructuring of the database that was undertaken last fall. You will need to upgrade to a compatible version of PhamNexus. I can see that Christian in the Hatfull lab is still keeping it up to date as he posted code changes last October to make it compatible with the new database format. He will need to tell you how to update to that working version.

Posted in: Bioinformatic Tools and Analyses → PhamNexus on SEA-VM

Link to this post \| posted 09 Feb, 2021 19:33
cdshaffer	This was due to a database versioning issue (i.e. starterator was still on version 391 and phagesdb had upated to version 392). The report is now available here. If you want all the details on the how and why this happens, and how to check versions see this thread: this thread If you want to see which version Phamerator.org, pecaan and starterator are on use these instructions: phagesdb see: http://databases.hatfull.org/Actino_Draft/Actino_Draft.version starterator see: http://phages.wustl.edu/starterator/database.version pecaan: look on any "pham maps" page just above the map phamerator.org: open the pull down menu in the top left

Link to this post | posted 09 Feb, 2021 19:33

cdshaffer

This was due to a database versioning issue (i.e. starterator was still on version 391 and phagesdb had upated to version 392). The report is now available here. If you want all the details on the how and why this happens, and how to check versions see this thread: this thread

If you want to see which version Phamerator.org, pecaan and starterator are on use these instructions:
phagesdb see: http://databases.hatfull.org/Actino_Draft/Actino_Draft.version
starterator see: http://phages.wustl.edu/starterator/database.version
pecaan: look on any "pham maps" page just above the map
phamerator.org: open the pull down menu in the top left

Posted in: Starterator → Pham not found in Starterator

Link to this post \| posted 23 Jan, 2021 19:33
cdshaffer	When it comes to enzyme naming issues I prefer the KEGG and exapasy "Enzyme" databases. For example here is the info on this enzyme at Expasy: https://enzyme.expasy.org/EC/2.7.6.1 The preferred name is Ribose-phosphate diphosphokinase, so they agree with your preference for "kinase". I also note that one of the mentioned synonyms also uses kinase. I have always been pretty agnostic on which term we should put on our approved function list but once it is there I think it makes all our work better if we all try to use that particular term as it creates a more consistent data set across all the SEA phages. phosphoribosyl transferase appears to be a general term as it hits to about 25 different enzymes in the Expasy database, the vast marjority of which are glycosyltransferases.

Link to this post | posted 23 Jan, 2021 19:33

cdshaffer

When it comes to enzyme naming issues I prefer the KEGG and exapasy "Enzyme" databases. For example here is the info on this enzyme at Expasy:

https://enzyme.expasy.org/EC/2.7.6.1

The preferred name is Ribose-phosphate diphosphokinase, so they agree with your preference for "kinase". I also note that one of the mentioned synonyms also uses kinase. I have always been pretty agnostic on which term we should put on our approved function list but once it is there I think it makes all our work better if we all try to use that particular term as it creates a more consistent data set across all the SEA phages.

phosphoribosyl transferase appears to be a general term as it hits to about 25 different enzymes in the Expasy database, the vast marjority of which are glycosyltransferases.

Posted in: Functional Annotation → Phosporibosylpyrophosphate synthetase

Link to this post \| posted 15 Dec, 2020 21:05
cdshaffer	ok release of starterator reports for database version 387 are now available and are tagged as the current version. Links from pecaan and phagesdb should now be working (at least until the next database update). Please report if any links fail at this point.

Posted in: Starterator → Read First: Common Starterator Troubleshooting

← previous
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
next →

Recent Activity

All posts created by cdshaffer