Posttranslational modifications

1
Welcome to Fortaleza
2
First a big welcome to the speakers

• We have a wonderful cast of speakers coming from
  all over the world
• Many of them have played, directly or indirectly,
  an important role in the history of Swiss-Prot
• In the name of the organizing committee I thank
  all of them for having accepted to participate to
  this anniversary meeting.

3
Then a big welcome to the UniProters

• Since 2000, every two years, the members of the
  Swiss-Prot groups at SIB and EBI have attended a
  retreat to discuss various aspects of their
  collaborations
• This meeting doubles up as a very special
  retreat. It is an opportunity for attendees and
  speakers to tell us what they think we should be
  doing!

4
Welcome and thank you to all attendees

• For those coming from all over the world we know
  it was not easy to come to Brazil
• We hope that this meeting will be an opportunity
  for you to listen to interesting talks
• But more importantly meetings are essential to
  network and to start or pursue collaborations
• So please enjoy and make the most of these four
  days that we will spend together.

5
And finally, but no the least! a big thank you
to all the sponsors
6
A few important last minute informations and
reminders

• In the conference bag external pocket you will
  find many important things including
• The pocket guide (the program at a glance)
• The instructions to access the Wifi Internet
  wireless service
• The vote bulletin for the best poster award
• Information about a survey concerning
  UniProtKB/Swiss-Prot.

7
The Swiss-Prot survey

• The Swiss-Prot annotators that are carrying out
  the survey have a small red sticker on their name
  badge
• The persons that will have answered the survey
  will receive a small yellow sticker to put on
  their name badge so that they do not get asked to
  participate over and over again!

8
Protein Spotlight book

• In your bag you will find a copy of Tales from a
  small world
• It is a book containing all the Protein Spotlight
  articles published since 2000
• We can offer a copy of this book to all of you
  thanks to Current Biodata who fully sponsored the
  cost of its printing.

9
Program changes

• Due to flight problems (Varig!) we lost 3
  speakers Terri Attwood, Philipp Bucher and
  Minoru Kanehisa
• We will use 2 of the 3 slots for different
  tutorials on Swiss-Prot, the third slot will be
  used to get to the beach an hour earlier on
  tuesday!
• Nasri Nahas talk will be given by Ron Appel as
  Nasri is busy trying to get his family out of
  Lebanon (last minute they are safely back in
  Geneva)
• Vitek Tracz talk will be given by Matthew
  Cockerill who has overall responsibility for
  BioMed central
• Really last minute we just learned Gunnar fell
  sick on the way here and has turned back to
  Sweden.

10
Speakers

• Try to end your talk 5 minutes before the alloted
  time slot so as to leave the opportunity for a
  few questions
• There is for each day of the conference a
  Swiss-Prot team member who is responsible for
  making sure we are on time and to moderate the
  question session
• You will have in front of you a digital timer
  that will show you how much time is left
• Once your time is over, it will ring and the
  moderator will make his best efforts to expel you
  from the podium!

11
Speakers - 2

• You can use the podium microphones and this will
  ensure that your image is captured correctly on
  the camera, but you can also use a wireless lapel
  mike
• Please use the mouse to point on objects in your
  presentation instead of the laser pointer.

12
The SwissProt Song Genome annotators, with your
big machines If you didn't have Swiss-Prot, You
wouldn't find a thing - with your big machines
- It you didn't have Swiss-Prot you would not
find a thing Ain't no good the software the grid
and the middleware If it'not for Swiss-Prot You
wouldn't get nowhere - with your middleware - If
it's not for Swiss-Prot You would not get nowhere
13
"Plus ça change, plus c'est la même chose the
next 20 years
14
the (pre)-history of Swiss-Prot
This will not be a talk on
15
The universe in which Swiss-Prot evolves
1953 1st sequence (bovine insulin) 1986 4000
sequences 2006 3.5 million sequences
Where will it stop?
179'000'025'042 (179 billion)
16
179'000'025'042
1st estimate 30 million species (1.5 million
named)
2nd estimate 20 million bacteria/archea
x 4'000 genes 5 million
protists x 6'000
genes 3 million insects
x 14'000 genes 1 million fungi
x 6'000 genes 0.6
million plants x
20'000 genes 0.2 million molluscs, worms,
arachnids, etc. x 20'000 genes 0.2 million
vertebrates x 25'000
genes
The calculation 2x107x40005x106x60003x106x14000
106x60006x105x200002x105x200002x105x25000
25000(Craig Venter)42(Douglas Adam)
Caveat this is an estimate of the number of
potential sequence entries, but not that of the
number of distinct protein entities in the
biosphere.
17
When will UniProtKB be complete?

• Swiss-Prot
• In July 2009 500000 entries
• In 2013 1 million entries
• In 2026 (40th anniversary) 10 million entries
• In 2036 (50th anniversary) 100 million entries.
• TrEMBL
• In May 2080 TrEMBL will have reached 10 billion
  entries
• We cant compute with Excel when we will reach
  179 billion entries
• But we are confident these dates are worthless as
  new sequencing techniques will have made all of
  these projections a very futile exercise!

18
Sequences

• The bread of Swiss-Prot. And yes annotations are
  the butter!
• gt99 of the protein sequences originate from
  translation of mRNA or genomic sequences
• Do we still need manual intervention to cater for
  sequences or can we just build smart filters to
  obtain those we want from TrEMBL?

19
So what is the current status?

• A snapshot of the situation
• 28200 entries with 82000 sequence conflicts
• 2600 entries with corrected frameshifts
• 15100 entries with corrected initiation sites
• 4300 entries with other sequence problems.
• At least 43000 entries (19 of Swiss-Prot)
  required a minimal amount of curation effort so
  as to obtain the correct sequence.

20
Quality of protein information from genome
projects

• Lets look at proteins originating from 3
  different genome projects
• Drosophila the example of what a curated (thanks
  to FlyBase) genome effort should look like only
  1.8 of the gene models conflict with what we
  have in Swiss-Prot
• Arabidopsis a typical example of a genome where
  lots of work was spent to annotate it at the time
  where it was sequenced, but where nothing as been
  done since (at least in the public view) 19.5
  of the gene models are erroneous
• Tetraodon nigroviridis the typical example of a
  quick and dirty automatic run through a genome
  with no manual intervention gt90 of the gene
  models produce incorrect proteins.

21
Human sequence entries as an example

• We have about 14500 human entries in Swiss-Prot
• 4300 entries contain information about 8000
  splice variants
• 4600 entries contain information about 27000
  sequence variants
• 7500 entries contain information about 22000
  sequence conflicts
• In average each human entry is produced by
  merging together sequence information from 6.2
  different nucleotide sequence entries.

22
Take home message

• Producing a clean set of sequences is not a
  trivial task
• It is not getting easier as more and more type of
  sequence data gets submitted
• It is important to pursue our efforts in making
  sure we provide to our users the most correct set
  of sequences for a given organism.

23
Post-translational modifications (PTMs)

• If sequences are important, their are generally
  not fully representative of the final biological
  entity most proteins are the target of PTMs
• PTMs are important at various levels, including
  the 3D structure, interactions, subcellular
  location and also the function
• The story of the integration of PTMs in
  Swiss-Prot consists of 3 distinct parts
• 1st part a long time ago in a distant
  proteogalaxy

FT MOD_RES 86 86
GAMMA-CARBOXYGLUTAMIC ACID. FT MOD_RES 110
110 HYDROXYLATION. FT CARBOHYD 203
203 POSSIBLE.
24
The 2nd phase 2000 to 2005

• Complete overhaul and significant extension of a
  controlled vocabulary for PTMs
• Creation of a PTM annotation program within the
  Swiss-Prot groups at SIB and EBI
• Development of new tools (Sulfinator, DGPI) for
  the prediction of some PTMs
• Massive clean up and re-annotation of many
  classes of PTMs.

25
The expanding world of PTMs

• We now have 283 different PTM descriptions
  (excluding processing, disulfide bonds and
  glycosylation events).

26
The new document listing post-translational
modifications
Contains many information items and is available
in html format or by ftp in tab-delimited format.
27
Finally LSEs for PTMs!

• Finally Proteoman has arrived! And PTM
  information can now be obtained from results of
  proteomics large scale experiments (LSE)
• In the past 12 months we have added about 6000
  experimental PTMs using data originating from
  some of these projects.

28
But LSEs are not so easy to deal with

• Issues mundane to the incorporation of LSE PTM
  data
• Quality
• Trying to assess whether the methodology really
  allows the detection of in- vivo modifications
• How many false positives are expected (often
  absent or very well hidden!)
• Accessing the data
• Often in supplementary material tables and in a
  variety of formats (HTML tables, excel
  spreadsheets, etc.)
• With a variety of identifiers (UniPRotKB, NCBI
  gi, pID, etc.)
• Sanity checking
• Making sure that the right sequence position is
  modified
• Does it make sense in the biological context
• Propagating the information to orthologs.
• So the big issue is how will we be able to scale
  up and deal with the expected increase in the
  number of such projects!

29
Cross-references then

• The DR lines were introduced in release 4 in
  April 1987 they first linked Swiss-Prot to EMBL,
  PDB and PIR
• They were instrumental in the development of SRS
  by Thure Etzold in the early 90s
• And also for ExPASy, the first web server in the
  life sciences in 1993.

30
Organism-specific gene databases AGD DictyBase Ech
oBASE EcoGene FlyBase GeneDB_Spombe GeneFarm Grame
ne HGNC H-InvDB HIV LegioList Leproma ListiList M
aizeDB MGI MIM MypuList PhotoList RGD SagaList SGD
StyGene SubtiList TAIR TubercuList WormBase WormP
ep ZFIN
Family and domain databases Gene3D HAMAP InterPro
PANTHER PIRSF Pfam PRINTS ProDom PROSITE SMART TIG
RFAMs
Enzyme and pathway databases BioCyc Reactome
Sequence databases EMBL PIR UniGene
2D-gel databases ANU-2DPAGE Aarhus/Ghent-2DPAGE C
OMPLUYEAST-2DPAGE ECO2DBASE HSC-2DPAGE OGP PHCI-2D
PAGE PMMA-2DPAGE Rat-heart-2DPAGE Siena-2DPAGE SWI
SS-2DPAGE
UniProtKB/Swiss-Prot explicit links
Protein family/group databases PptaseDB GermOnline
MEROPS REBASE TRANSFAC
Miscellaneous dbSNP GO IntAct LinkHub RZPD-ProtExp

3D structure databases HSSP PDB SMR
Genome annotation databases Ensembl GenomeReviews
TIGR
PTM databases GlycoSuiteDB PhosSite
31
Cross-references now

• There are now cross-references from Swiss-Prot to
  74 different databases (6 more are in the
  pipeline)
• Almost 3 million DR lines an average of 12 per
  entry
• Many other links to external resources are also
  available through the OX (NCBI taxonomy), RX
  (PubMed, DOI), CC (Web resource topic) and FT
  lines (dbSNP)
• Cross-references are not only a mean to help
  navigate between resources, they sometimes add
  information to the entries.

32
Examples of cross-references that provide
information

• The cross-references to the Gene Ontology (GO)
• DR GO GO0005634 Cnucleus ISS.
• DR GO GO0005515 Fprotein binding IPI.
• DR GO GO0007165 Psignal transduction TAS.
• The PDB cross-references include information on
  the mapping of the structure on the sequence
• DR PDB 1QQG X-ray A/B4-267.
• The cross-references to domain databases include
  information on the name/acronyms of the domains
  and the number of occurrences of these domains
• DR PROSITE PS50026 EGF_3 2.
• DR PROSITE PS50092 TSP1 3.
• DR PROSITE PS01208 VWFC_1 1.

33
From sequences to structures..and back!

• Efficient bidirectional links between UniProtKB
  and PDB/MSD are very important
• Currently 10000 Swiss-Prot entries are linked to
  30200 PDB entries
• These links are constantly updated and verified
  the converse is unfortunatly still not yet true
• We have always made use of 3D structure
  information to help in the annotation process
• But we are only now starting to systematically
  mine 3D structures to extract various information
  such as disulfide bonds, metal-binding sites,
  active sites, etc.

34
So what is the future of cross-references?

• Will we really need hard-coded cross-references
  in the future?
• Can we gradually replace some of them by computed
  on the fly links using referenceable objects?
• Will we make more use of client-server systems
  such as the distributed annotation system (DAS)?
• The answer is obviously dependent on
  standardization
• But the Life Sciences are still living in the
  dark ages of the tower of Babel

35
CVs and ontologies

• Since the very beginning of Swiss-Prot we have
  been building a growing sets of controlled
  vocabularies (ontologies)
• Species, strains, plasmids, journals, tissues,
  PTMs domain names and, of course, keywords are
  all under control (see posters SP117 and
  SP120)
• We are very well advanced in the process of
  having a CV for pathways (see the UniPathway
  poster SP140)
• We are now tackling the problems of protein and
  gene names (see poster SP118). But this is of
  course not very easy!

36
Do we need annotations?

• Annotators spend a big part of their time
  capturing and synthesizing a huge amount of
  functional information
• For example we populate Swiss-Prot with data
  relevant to the
• Role and function of the proteins
• Subcellular location
• Interactions (binary and complex)
• Tissue specificity, developmental stage
• Involvement in diseases.
• We have many anecdotal evidence that users find
  this very important and that this is one of the
  important hallmark of Swiss-Prot. Yet is this
  really true?

37
Do we need annotations? part 2

• This is a time consuming process and we will
  never be complete and up-to-date
• Many users want quick and easy to summarize
  answers, yet the more detailed an entry becomes
  the less it is easy to transform it into a
  summarizable entity
• We are often the victims of the fasta format
  syndrome users expect everything important
  about a protein to be available in the header of
  a fasta format entry!
• So should we continue?

38
Yes we need annotation!

• Because (among many other reasons)
• Automatized annotation is the only way to
  transfer knowledge from a model organism to a
  less studied one
• To apply such techniques safely one needs
  template entries that are representative of the
  state of the knowledge
• While literature mining tools could be conceived
  as a way to automatically build a summary view of
  the knowledge around a given protein, these
  techniques are not yet powerful enough to create
  a coherent synthetic view
• Literature mining tools also require the
  existence of well annotated (corpus) entries.

39
From pull to push..

• For now more than 20 years we have been pulling
  information and knowledge from various sources,
  but mainly from literature
• It is now time to make sure that the next 20
  years will be defined by the fact that
  researchers push their results and the
  interpretation of their results in the
  knowledgebase.

40
Adopt a protein

• Attempt to try to get the community to directly
  submit information on the proteins that they are
  studying
• Using a wikepedia-type model/interface
• Will first be field-tested in the yeast
  community
• We are hopeful, yet we are realist only a small
  percentage of life researchers will take the time
  and are altruistic enough to fully participate in
  such a scheme.

41
Grey grey matter counts!

• Many life scientists with knowledge of the
  molecular world and that are computer-proficient
  are reaching retirement age
• Some want to continue to play a role in the
  advancement of research, yet they will not be
  able to do lab work anymore
• We should offer them the tools necessary for them
  to contribute to the annotation process.

42
Anabelle and Asterix

• Two important tools could contribute to the
  democratization of Swiss-Prot style annotation
• Anabelle a web based protein sequence analysis
  platform
• Asterix the new Swiss-Prot editor.

43
Anabelle selection module
Viewer Layout
Link to entry NiceProt view
Blast (full) entry
more links!...
Links
Link to InterPro
Link to domain original database
Link to most similar entry NiceProt view
Align most similar entry with entry
Blast uncharted region
44
And here is what the users gets back
45
But what about the rest of the life scientists?

• We saw how we could get parents (adopt a protein)
  and grand parents (grey matter count)
  involvements, but what about the children..
• the young researchers, those who are active in
  producing new knowledge?

46
Two carrots, a stick and lots of education!

• The carrots
• Making sure that granting agencies see favorably
  the involvement of researchers in the process of
  submitting information to databases
• The same criteria should be considered by any
  hiring or promotion committee
• The stick getting journal editors to refuse to
  accept to publish a paper if the results have not
  been submitted to the relevant knowledge
  resources

47
Education!

• Everyone should feel concerned
• Awareness of the content and usage of knowledge
  resources is a pre-requisite to do any type of
   serious  research in the field of molecular
  life sciences
• Organizations such as EMBNet, EBI, SIB, NCBI, NIG
  should continue and strenghten their outreach
  efforts
• We (databases providers) should do more in term
  of providing tutorials (on-line and on-site).

48
An important issue

• The process of developing a data resource for the
  Life Sciences is akin to the work of middle age
  copists, renaissance encyclopedists or the 19th
  century OED development it is a very tedious,
  manually intensive, long term job

49
How to get funding for knowledge infrastructures
in the life sciences?

• Funding knowledge resources is difficult
• Its a very long term process
• Its not prestigious
• and its not cheap!

50
And its not only databases that are endangered!
Service groups are also at risk
51
Proposition for a new tax

• Each grant proposal for a high throughput
  data-producing project would be obliged to set
  aside a predefined percentage of the grant money
  to help cover the cost of storing and managing
  the produced data
• How this money would be redistributed is not
  trivial to define and even less to implement
• The priority would be to use this tax as a
  financial tool to help fund the data repositories.

52
The 6 observations of a  databaser 

• Your task will be much more complex and far
  bigger that you ever thought it could be
• If your database is successful and useful to the
  user community, then you will have to dedicate
  all your efforts to develop it for a much longer
  period of time than you would have thought
  possible
• You will always wonder why life scientists abhor
  complying with nomenclature guidelines or
  standardization efforts that would simplify your
  and their life
• You will have to continually fight to obtain a
  minimal amount of funding
• As with any service efforts, you will be told far
  more what you do wrong rather than what you do
  right
• But when you will see how useful your efforts are
  to your users, all the above drawbacks will loose
  their importance!!

53
Thank You
Aiala, Alain x4, Alan x4, Alastair, Alex x2,
Alexander x2, Alexandre x2, Alice, Alistair,
Allyson, Alvis, Amanda, Ana Tereza, Anastasia,
Andre x3, Andrea, Andreas, Andrew, Angela, Anne
x4, Anne-Lise, Anthony, Antoine, Anulka, Arnaud
x2, Arthur, Astrid, Athel, Barbara x2, Barend,
Baris, Barry, Bart, Bastien, Bengt, Bernard x2,
Bernd, Bernhard x2, Bill, Bob, Brigitte, Bruno
x2, Burkhard, Carl, Carola, Carolyn, Catherine
x4, Cathy x2, Cecile x2, Cecilia, Cedric, Cesare,
Chantal x3, Charles x2, Chris, Chrissie,
Christian x3, Christiane, Christine x2,
Christoph, Christophe, Christopher, Christos,
Claude, Claudia x2, Claudine, Colin, Colombe,
Corinne, Cristiano, Damien, Dan, Dana, Daniel x3,
Daniela, Danielle, Darcy, Darren, Dave x2, David
x5, Delphine, Denis x2, Dennis, Des, Dietmar,
Dolnide, Dominique, Doron, Dorothy, Doug, Duncan,
Eddie, Edgar, Edouard, Eleanor, Elisabeth x2,
Elmar, Elvis, Emily, Emmanuel, Eric x3, Erik,
Ernest, Ernst, Esther, Eugene x2, Eva, Eve,
Evelyn, Evgenia, Evgeny, Ewan, Fabrice, Fiona,
Flavio, Florence x3, Fotis, Francis, Frank,
François x3, Frederic, Frederique x2, Gabriel,
Gabriella, Ganesh, Gaston, Geoff, Gerry, Gert,
Ghislaine, Gilbert, Gill, Goran, Gottfried,
Graham x2, Greg, Gregoire, Guido, Guillaume,
Gunnar, Guy x2, Guy-Olivier, Hanah, Heidi,
Henning, Hien, Hilde, Holger, Hongzhan, Howard,
Hsing-Kuo, Ian, Iirit, Ilkka, Ioannis, Irving,
Isabelle x2, Ivan x2, Ivo, Jack, Jacques x2,
Jaime, Janet x2, Jean-Charles, Jean-François,
Jean-Jacques, Jean-Michel, Jean-Pierre x2,
Jeffrey, Jenny, Jerome, Jim, Jingchu, Joachim,
Joanna, Joel, John x7, Jonas, Jonathan x2, Jorja,
Jos, Juan, Juergen, Julia, Julio, Julius, Kai,
Karin, Karine, Kate, Kati, Katja, Katsumi, Kay,
Keiichi, Keith x3, Ken x2, Kenta, Khaled, Kirill,
Kirsty, Kristian, Larry, Laure, Laurent x3, Lee,
Leigh, Leon, Li, Lina, Lionel, Lisa x2, Livia,
Lorenza, Lorenzo, Louise, Luca, Luciane, Lucien,
Luisa x2, Luiz, Lydie x2, Ma'ayan, Madelaine,
Maggie, Mahesh, Manolo, Manuel x2, Manuela, Marc
x6, Marcia, Marco, Margaret x2, Mari Trini,
Maria, Maria Esperanza, Maria-Jesus,
Marie-Claude, Marilyn, Marisa, Mark x2, Martin
x2, Martine, Marvin, Mary, Massimo, Matteo,
Matthew, Mauricio, Michael x7, Michel x3,
Michele, Michelle, Miguel, Mike x2, Minna,
Minoru, Monica, Monika, Morido, Nabil, Nadeem,
Nadine x2, Naruya, Nasri, Natalia, Nathalie, Neil
x2, Nicky, Nicola, Nicolas x3, Nicole x3,
Nicoletta, Nicolle, Nikos, Nina, Oliver, Olivier
x4, Orna, Owen, Paolo, Pascal, Pat, Patricia x6,
Patrick x5, Paul, Paula, Pavel, Pedro, Peer,
Peter x7, Petra, Phil x2, Philip, Philippe x3,
Pierre, Pierre-Alain, Pieter, Piotr, Rachael,
Raffaella, Rainer, Raja, Rasko, Raton laveur,
Rebecca x2, Rein, Reinhard x2, Remi, Reto,
Reynaldo, Rich, Richard, Robert x2, Roberto,
Robin, Rodger, Rodrigo, Roland, Ron, Rosita,
Ross, Roy, Russ x2, Ruth x3, Saeid, Salvo, Samia,
Samuel x2, Sandor, Sandra x2, Sandrine, Sarah,
Scott, Sebastien x2, Serenella, Sergio, Severine
x2, Shigehaki, Shmuel, Shoko, Shoshana, Shyamala,
Silvia x2, Sineaid, Siv, Sona, Soren, Sorogini,
Steffen x2, Steffi, Stephanie x2, Steve, Steven,
Stuart x2, Stylianos, Sunil, Sylvain, Sylvie x2,
Takashi, Tamara, Tammera, Tania x2, Temple,
Terri, Terry, Thomas x3, Thure, Tim x2, Timothy,
Toby, Tom, Toni, Torsten, Ujwal, Ulrich, Ursula,
Valeria, Vassilios, Veronique, Vicente, Victor
x2, Vincent, Vinnei, Violaine, Virginie x2,
Vitaliano, Vitek, Vivien x2, Vivienne, Wanessa,
Wei mun, Weimin, William, Williams, Willy,
Winona, Winston, Witek, Wolfgang, Xavier x2,
Yasmin, Yasuhiro, Yongxing, Yoshio, Youla,
Young-Ki, Zeev, Zhang-Zhi.