July 2011 - Mg-rast - mailman3.cels.anl.gov

Re: [mg-rast] Doubt about MG-Rast annotation
by Gustavo B. Gregoracci 19 Jul '11

19 Jul '11

Hi there! I'm not sure I got it right. Let's see... I annotated the metagenome through Genbank, for example. The "# proteins" field gives me the number of unique sequences which matched the genbank database and the "abundance" is the total number of matches including repeated matches to the "# proteins" proteins. Is that so? When I send these to the workbench the "# proteins" is sent; and this subset is re-annotated with Subsystems, for example. The "#proteins" falls since not all of the original set annotated with Genbank will be recognized here, right? But the abundance may increase because more repeats can be found in the larger database? And what the "workbench abundance" means? Sorry to bother you further, but I'm could not grasp the explanation previously, Thanks again for the help and patience, Best regards, Gustavo Gustavo Bueno Gregoracci -- MSc., Dr. in Microbiology (Genetics and Molecular Biology) Email: gustavo_biomed(a)yahoo.com -- Pos doc Student Laboratório de Microbiologia (Prof. Fabiano Thompson) Centro de Ciências de Saúde - Instituto de Biologia - UFRJ Tel: +55 (21) 2562-6567 >________________________________ >From: Mark DSouza <dsouza(a)mcs.anl.gov> >To: Gustavo B. Gregoracci <gustavo_biomed(a)yahoo.com> >Cc: MG- Rast <mg-rast(a)mcs.anl.gov>; Louisi Oliveira <louisioliveira(a)gmail.com> >Sent: Monday, July 18, 2011 7:13 PM >Subject: Re: Doubt about MG-Rast annotation > >Hi, > >The number discrepancy arises from the use of different databases, the original 209 hits is specific to the SEED subsystem, with 66 hits against the proteins in these organisms. When you perform the organism classification against GenBank, the number of hits is higher because of the greater number of organisms in the GenBank database, and so does the abundance counts. The workbench abundance column of the table gives a better breakdown of the taxonomic counts as compared to the 209 count. This is not very intuitive from the display and we are looking into changing the table to make this more explicit. > >Regards, >Mark > >----- Original Message ----- >> From: "Gustavo B. Gregoracci" <gustavo_biomed(a)yahoo.com> >> To: "Mark DSouza" <dsouza(a)mcs.anl.gov> >> Cc: "MG- Rast" <mg-rast(a)mcs.anl.gov>, "Louisi Oliveira" <louisioliveira(a)gmail.com> >> Sent: Monday, July 18, 2011 1:03:28 PM >> Subject: Re: Doubt about MG-Rast annotation >> Hi there Mark, >> >> >> Thanks for the reply. So, the metagenome ID is 4465448.3 and its >> called Forno. The glitch has actually happened in others as well, and >> I'm only using this one as an example. >> >> >> I re-checked the numbers and they still don't add. I'm working with >> Genbank (e-5) for organism classification, and subsystems (e-5) for >> functional classification, for that matter. >> >> >> I have 209 hits in this entire metagenome for photosynthesis. If I >> check the number of photosynthesis hits among the bacterial hits I get >> 140, while the same for eukaryotic gives me 142 hits. That puzzled me. >> When I do the opposite, and check the taxonomic composition of just >> the 209 photosynthesis hits the numbers are even more weirder. I get >> 245 hits for bacteria and 403 for eukaryotes! >> >> >> Thanks again for the help, >> I'm available for any further explanations necessary, >> >> >> Best regards, >> Gustavo >> >> Gustavo Bueno Gregoracci >> -- >> MSc., Dr. in Microbiology (Genetics and Molecular Biology) >> Email: gustavo_biomed(a)yahoo.com >> >> -- >> Pos doc Student >> Laboratório de Microbiologia (Prof. Fabiano Thompson) >> Centro de Ciências de Saúde - Instituto de Biologia - UFRJ >> Tel: +55 (21) 2562-6567 >> >> >> >> >> >> From: Mark DSouza <dsouza(a)mcs.anl.gov> >> To: Gustavo B. Gregoracci <gustavo_biomed(a)yahoo.com> >> Cc: MG- Rast <mg-rast(a)mcs.anl.gov> >> Sent: Monday, July 18, 2011 1:33 PM >> Subject: Re: Doubt about MG-Rast annotation >> >> Hi, >> >> Can you please send us the MG-RAST ID for the dataset where you >> observed this, we will check into it for you. >> >> -- >> Regards, >> Mark D'Souza >> -- for the MG-RAST team >> >> mg-rast(a)mcs.anl.gov >> http://metagenomics.anl.gov/ >> >> If you need to respond to this email please address it to the >> mailing-list mg-rast(a)mcs.anl.gov >> >> >> ----- Original Message ----- >> > From: "Gustavo B. Gregoracci" < gustavo_biomed(a)yahoo.com > >> > To: "Mark D'Souza" < dsouza(a)mcs.anl.gov > >> > Sent: Saturday, July 16, 2011 10:05:14 AM >> > Subject: Doubt about MG-Rast annotation >> > Hello there Mark, >> > >> > >> > >> > As I was analyzing some metagenomes, I ran into a recurrent problem. >> > I >> > was trying to cross information from taxonomic and functional >> > annotation, and discovered some inconsistencies. I could not figure >> > those out so I’m writing you in the hope that you can help me, since >> > this will affect my analysis. >> > >> > >> > You see… I wanted to understand the contribution of different >> > phylogenetic groups to a given subsystem. So I got the total number >> > of >> > sequences for the photosynthesis subsystem, for example, which was >> > 209 >> > hits. Then I went into taxonomy and separated all entries identified >> > as eukaryotes, sending them to the workbench. I changed again to >> > functional and checked the total number of eukaryotic hits to >> > photosynthesis, which was 163. So far so good, but I decided to >> > perform the same analysis regarding prokaryotes. Sent all >> > prokaryotic >> > entries to the workbench and changed to functional to see their >> > contribution to photosynthesis as well. I got 174 hits! >> > >> > >> > >> > How is this possible? If my whole metagenome has 209 hits to a >> > subsystem, how can I have 163 eukaryotic hits within it and 174 >> > prokaryotic hits to the same subsystem? Can sequences be annotated >> > as >> > both eukaryotic and prokaryotic? Cause I was interpreting these >> > categories as mutually exclusive… >> > >> > >> > Thanks in advance for the help, >> > Looking forward to hear from you, >> > >> > >> > Best regards, >> > Gustavo >> > >> > Gustavo Bueno Gregoracci >> > -- >> > MSc., Dr. in Microbiology (Genetics and Molecular Biology) >> > Email: gustavo_biomed(a)yahoo.com >> > >> > -- >> > Pos doc Student >> > Laboratório de Microbiologia (Prof. Fabiano Thompson) >> > Centro de Ciências de Saúde - Instituto de Biologia - UFRJ >> > Tel: +55 (21) 2562-6567 > > >

1 0

Re: [mg-rast] PCA
by Folker Meyer 19 Jul '11

19 Jul '11

Ella try a screenshot instead? if you are on a mac that is very easy. there is the grab tool pre-installed on your mac... On Jul 19, 2011, at 6:43 PM, Ella Rakowski wrote: > Hi- > > I am trying to create a PCA from a few metagenomes, and when I click on the button to make it static (so I can save the image and use it) the image disappears. The grouping info, r^2 values, raw data plots, etc are still there. > > Thanks! > > Ella Rakowski

1 0

PCA
by Ella Rakowski 19 Jul '11

19 Jul '11

Hi- I am trying to create a PCA from a few metagenomes, and when I click on the button to make it static (so I can save the image and use it) the image disappears. The grouping info, r^2 values, raw data plots, etc are still there. Thanks! Ella Rakowski

1 0

Fwd: bug for tree generation
by Mark DSouza 19 Jul '11

19 Jul '11

forward ----- Forwarded Message ----- > From: amarots(a)iqsc.usp.br > To: "Mark D'Souza" <dsouza(a)mcs.anl.gov> > Sent: Tuesday, July 19, 2011 4:18:00 PM > Subject: Re: [mg-rast] bug for tree generation > Ok, > Thanks a lot Mark. > regards, > Amaro > > > > Hi, > > > > The database query required to create the trees becomes more complex > > with > > number of metagenomes and increasing number of filtering steps > > causing > > lengthy times for retrieval of data and may cause the tree > > visualization > > to fail, we will check and see what can be done. > > > > Regards, > > Mark D'Souza > > -- for the MG-RAST team > > > > mg-rast(a)mcs.anl.gov > > http://metagenomics.anl.gov/ > > > > If you need to respond to this email please address it to the > > mailing-list > > mg-rast(a)mcs.anl.gov > > > > > > > > > > On Jul 19, 2011, at 9:03 AM, amarots(a)iqsc.usp.br wrote: > > > >> Hello MG-RAST team, > >> I wonder if you could give me a help. I can't visualize a specific > >> data > >> set saved at workbench as a tree scheme. I picked metagenomes: > >> 4461456.3, > >> 4461455.3, 4461454.3, 4461453.3,4461452.3, 4461451.3, 4453374.3, > >> 4461444.3, and pooled the Bacteria domain related proteins (total > >> of > >> 48557 > >> unique proteins, using GenBank and cutoff of 1e-5 and Min Alig > >> Length of > >> 50bp). When I try to generate a tree using this set of ptns the > >> "data > >> generation progress" keep going forever. When I use all ptns info > >> from > >> this group of metagemones, a shorter workbench set of ptns or less > >> metagenomes I could get the tree scheme tough. So, it seems to be a > >> bug > >> that happens specificaly with the set of ptns that I chose. > >> Thanks, > >> Amaro > >> > >> > >> > >> > > > >

1 0

Re: [mg-rast] Lowest Functional Classification
by Mark D'Souza 19 Jul '11

19 Jul '11

Hi, Sorry, we do not have a comprehensive list. The information is available from the individual databases though, see the Annotation sources page at: http://metagenomics.anl.gov/metagenomics.cgi?page=Sources and click on each database to get download information. Regards, Mark D'Souza -- for the MG-RAST team mg-rast(a)mcs.anl.gov http://metagenomics.anl.gov/ If you need to respond to this email please address it to the mailing-list mg-rast(a)mcs.anl.gov On Jul 19, 2011, at 2:53 PM, j.matthewhaggerty(a)yahoo.com wrote: > Hi, I am a doctoral student at San Diego State University. I have been using MG-RAST to look at a number of marine metagenomes. I was wondering if there is a way to access a complete list of possible functional matches, not just the enzymes or genes my metagenomes hit. Ideally, I would also like to know which functional genes fall into multiple pathways so that I can account for sequences being counted multiple times. I know some of this information is available in SEED but a complete list would be very helpful. > Thanks for any help you can send my way, > J. Matthew Haggerty

1 0

Lowest Functional Classification
by j.matthewhaggerty＠yahoo.com 19 Jul '11

19 Jul '11

Hi, I am a doctoral student at San Diego State University. I have been using MG-RAST to look at a number of marine metagenomes. I was wondering if there is a way to access a complete list of possible functional matches, not just the enzymes or genes my metagenomes hit. Ideally, I would also like to know which functional genes fall into multiple pathways so that I can account for sequences being counted multiple times. I know some of this information is available in SEED but a complete list would be very helpful. Thanks for any help you can send my way, J. Matthew Haggerty

1 0

Re: [mg-rast] QIIME Report in MG-RAST v3.1
by Mark D'Souza 19 Jul '11

19 Jul '11

Hi Deepak, The data file is formatted as input for QIIME, with the md5sum for the sequence and assigned taxonomy. We will be adding an entry in the MG-RAST documentation regarding the use of these files, till then you can check out the QIIME documentation: http://qiime.sourceforge.net/tutorials/tutorial.html Regards, Mark D'Souza -- for the MG-RAST team mg-rast(a)mcs.anl.gov http://metagenomics.anl.gov/ If you need to respond to this email please address it to the mailing-list mg-rast(a)mcs.anl.gov On Jul 18, 2011, at 11:13 AM, Deepak Sharma wrote: > Hi Mark, > I hope you are doing great. I have a small question regarding the QIIME report feature in MGRAST. MGRAST website doesn't give much information about the QIIME feature like how to use the file and what the two files means. Could you give me some information about how *data and *metadata can be used by QIIME? > > Thanks > > -- > Deepak Sharma, M.Sc > Microbial Research Tech. > West Virginia State University > PO Box 1000 - 102 Hamblin Hall > Institute, WV 25112 > Telephone: (304) 766-5239 > E-mail: dsharma(a)wvstateu.edu >

1 0

Re: [mg-rast] taxonomic classification in mg rast 3
by Mark DSouza 19 Jul '11

19 Jul '11

I have some text, with some help we can add details and blog it. send me the diagrams/docs which can be used. m ----- Original Message ----- > From: "Folker Meyer" <folker(a)anl.gov> > To: "Mark D'Souza" <dsouza(a)mcs.anl.gov> > Cc: "MG- Rast" <mg-rast(a)mcs.anl.gov> > Sent: Tuesday, July 19, 2011 12:15:59 PM > Subject: Re: [mg-rast] taxonomic classification in mg rast 3 > do we not have an algorithms page on the blog? > > if not lets write one. I have some diagrams and docs for the > paper/.... > > > On Jul 19, 2011, at 11:50 AM, Mark D'Souza wrote: > > > Hi, > > > > The taxonomic classification in MG-RAST is obtained from the > > sequence similarity results against the database selected under > > 'annotation sources'. > > > > Regards, > > Mark D'Souza > > -- for the MG-RAST team > > > > mg-rast(a)mcs.anl.gov > > http://metagenomics.anl.gov/ > > > > If you need to respond to this email please address it to the > > mailing-list mg-rast(a)mcs.anl.gov > > > > > > > > > > On Jul 19, 2011, at 9:02 AM, <shuang(a)mpi-bremen.de> > > <shuang(a)mpi-bremen.de> wrote: > > > >> Good afternoon, dear mg-rast maintainers. > >> > >> I have a question about the taxonomic classification in mg-rast. I > >> see this > >> feature in the result page of my metagenome but even after I > >> searched the > >> help, no information regarding the tool, method, algorithm and so > >> on. So I > >> would really want to know which tool or tools are mg rast is using > >> and are > >> there any metatools or so are employed. Thank you very much. > >> > >> > >> Sixing Huang > >

1 0

Re: [mg-rast] Fwd: bug for tree generation
by Mark DSouza 19 Jul '11

19 Jul '11

Thanks, m ----- Original Message ----- > From: "Travis Harrison" <teharrison(a)anl.gov> > To: "Mark D'Souza" <dsouza(a)mcs.anl.gov> > Cc: mg-rast(a)mcs.anl.gov > Sent: Tuesday, July 19, 2011 1:31:01 PM > Subject: Re: [mg-rast] Fwd: bug for tree generation > Not really. > > There is a query size database limitation when accessing data from > workbench md5 sets. There is a SQL statment of 'select <stuff> from > <job_table> where md5 in (list of workbench md5s)' going on at the > backend. Once the list is too large the time taken becomes too long > and we get into webserver timeouts. 50,000 md5s is too much. It kind > of depends on the size of the data table, but we start getting > timeouts somewhere between 20-40,000 md5s. We could split up the query > and get the results in parallel, but we would still get a timeout at > some point as data gets larger. > > -Travis > > > On Tue, Jul 19, 2011 at 12:06 PM, Mark D'Souza < dsouza(a)mcs.anl.gov > > wrote: > > > Hi Travis, any hope for this guy? > Mark > > > Begin forwarded message: > > > From: Mark D'Souza < dsouza(a)mcs.anl.gov > > > Date: July 19, 2011 12:05:00 PM CDT > > To: amarots(a)iqsc.usp.br > > Cc: mg-rast(a)mcs.anl.gov > > Subject: Re: [mg-rast] bug for tree generation > > > > > > > Hi, > > > > The database query required to create the trees becomes more complex > > with number of metagenomes and increasing number of filtering steps > > causing lengthy times for retrieval of data and may cause the tree > > visualization to fail, we will check and see what can be done. > > > > Regards, > > Mark D'Souza > > -- for the MG-RAST team > > > > mg-rast(a)mcs.anl.gov > > http://metagenomics.anl.gov/ > > > > If you need to respond to this email please address it to the > > mailing-list mg-rast(a)mcs.anl.gov > > > > > > > > > > On Jul 19, 2011, at 9:03 AM, amarots(a)iqsc.usp.br wrote: > > > >> Hello MG-RAST team, > >> I wonder if you could give me a help. I can't visualize a specific > >> data > >> set saved at workbench as a tree scheme. I picked metagenomes: > >> 4461456.3, > >> 4461455.3, 4461454.3, 4461453.3,4461452.3, 4461451.3, 4453374.3, > >> 4461444.3, and pooled the Bacteria domain related proteins (total > >> of 48557 > >> unique proteins, using GenBank and cutoff of 1e-5 and Min Alig > >> Length of > >> 50bp). When I try to generate a tree using this set of ptns the > >> "data > >> generation progress" keep going forever. When I use all ptns info > >> from > >> this group of metagemones, a shorter workbench set of ptns or less > >> metagenomes I could get the tree scheme tough. So, it seems to be a > >> bug > >> that happens specificaly with the set of ptns that I chose. > >> Thanks, > >> Amaro > >> > >> > >> > >> > >

1 0

Re: [mg-rast] Fwd: bug for tree generation
by Travis Harrison 19 Jul '11

19 Jul '11

Not really. There is a query size database limitation when accessing data from workbench md5 sets. There is a SQL statment of 'select <stuff> from <job_table> where md5 in (list of workbench md5s)' going on at the backend. Once the list is too large the time taken becomes too long and we get into webserver timeouts. 50,000 md5s is too much. It kind of depends on the size of the data table, but we start getting timeouts somewhere between 20-40,000 md5s. We could split up the query and get the results in parallel, but we would still get a timeout at some point as data gets larger. -Travis On Tue, Jul 19, 2011 at 12:06 PM, Mark D'Souza <dsouza(a)mcs.anl.gov> wrote: > Hi Travis, any hope for this guy? > Mark > > > Begin forwarded message: > > > From: Mark D'Souza <dsouza(a)mcs.anl.gov> > > Date: July 19, 2011 12:05:00 PM CDT > > To: amarots(a)iqsc.usp.br > > Cc: mg-rast(a)mcs.anl.gov > > Subject: Re: [mg-rast] bug for tree generation > > > > Hi, > > > > The database query required to create the trees becomes more complex with > number of metagenomes and increasing number of filtering steps causing > lengthy times for retrieval of data and may cause the tree visualization to > fail, we will check and see what can be done. > > > > Regards, > > Mark D'Souza > > -- for the MG-RAST team > > > > mg-rast(a)mcs.anl.gov > > http://metagenomics.anl.gov/ > > > > If you need to respond to this email please address it to the > mailing-list mg-rast(a)mcs.anl.gov > > > > > > > > > > On Jul 19, 2011, at 9:03 AM, amarots(a)iqsc.usp.br wrote: > > > >> Hello MG-RAST team, > >> I wonder if you could give me a help. I can't visualize a specific data > >> set saved at workbench as a tree scheme. I picked metagenomes: > 4461456.3, > >> 4461455.3, 4461454.3, 4461453.3,4461452.3, 4461451.3, 4453374.3, > >> 4461444.3, and pooled the Bacteria domain related proteins (total of > 48557 > >> unique proteins, using GenBank and cutoff of 1e-5 and Min Alig Length of > >> 50bp). When I try to generate a tree using this set of ptns the "data > >> generation progress" keep going forever. When I use all ptns info from > >> this group of metagemones, a shorter workbench set of ptns or less > >> metagenomes I could get the tree scheme tough. So, it seems to be a bug > >> that happens specificaly with the set of ptns that I chose. > >> Thanks, > >> Amaro > >> > >> > >> > >> > > > >

1 0