Serialized Thoughts...

Saving your data with Assemble 1.0.4 and S2S 2.0.3

2011-02-15T16:22:00.000+01:00

With the new versions of Assemble and S2S, the way to name the files storing the 2D, 3D and alignments data has been changed. The following screenshot shows you a 3D model saved with Assemble 1.0.3:

Each molecule is saved in a file whose name corresponds to a unique ID. If this model is opened with Assemble 1.0.4 and saved, you will get the following result:

Assemble 1.0.4 and S2S 2.0.3 use now a unique ID to name the files storing the 2D, 3D and alignments data. You can notice that the original files are kept. Starting from now, only the files with the unique ID will be used. We suggest you to delete the old versions manually. If not, Assemble or S2S will always ask you to choose between the two versions of your data (the old and the new one).

Open position in RNA 3D modeling (Valencia, Spain)

2011-02-10T09:16:00.000+01:00

I'm forwarding this announcement tweeted by Pedro Beltrao this morning. All the details here.

Assemble 1.0.4 and S2S 2.0.3 are available

2011-02-09T22:54:00.001+01:00

You can get them from the Download sections of the S2S or Assemble websites. You will find mainly:
- a fix to load the Rfam alignments formatted with the Stockholm format (S2S)
- a fix to merge several secondary structures (2D Model -> Import... in Assemble)
- a fix for the first save of your data under Windows (S2S and Assemble)
- a fix to refine a 3D model made with several RNA molecules (Assemble)
- any RNA molecule of an S2S alignment can be defined as a reference molecule
- an S2S alignment can be extended with molecules stored in a PDB file
- an S2S alignment can be constructed from an Assemble model

When Assemble 1.0.4 is launched for the first time, it will take some time to re-index all the 3D structures stored in the MyPDB library.

You can find all the details about the improvements and the bugs fixed in the CHANGELOG files provided with each tool.

We hope that you will enjoy these new versions.

The publication of Assemble is available in Bioinformatics

2010-06-21T10:13:00.021+02:00

We're pleased to inform you that Bioinformatics has made available our manuscript of Assemble. It is freely available as an open access publication. We would like to thank the reviewers who have appreciated the work done.

We also would like to thank the S2S-Assemble community for its help and support. As an advance access version, the full citation is:

Assemble: an interactive graphical tool to analyze and build RNA architectures at the 2D and 3D levels
Fabrice Jossinet; Thomas E Ludwig; Eric Westhof
Bioinformatics 2010; doi: 10.1093/bioinformatics/btq321

The automatic update system is fixed for Windows in Assemble 1.0.3 and S2S 2.0.2

2010-06-09T10:45:00.002+02:00

If you're using Windows, you have probably observed that the update system is not working. The updating process is running well, but when you restart your graphical tool, it is not updated.

To update themselves, S2S and Assemble download the new files with the same name than the file to replace, but with a "_tmp" suffix. Once the new files fully downloaded, the "_tmp" version replaces the old one. Unfortunately, Windows is locking files when S2S or Assemble are running. So, it is not possible to replace files like assemble.jar, s2s.jar or paradise.jar. This is the situation i had under Windows 7 when i updated Assemble 1.0 to the recent 1.0.2 version:

You can see that i still have the new version as an assemble.jar_tmp file.

I fixed this bug in Assemble 1.0.3 and S2S 2.0.2. These versions are available today as automatic updates (Help Menu -> Update S2S/Assemble).

So, if you're using Windows, relaunch your update. This will make two modifications in your S2S/Assemble directory:
- a new file "updater.jar" appears
- the file launch_S2S/Assemble_for_Windows.bat has been improved. Don't forget to edit it with your favorite text editor to precise the location of your Java environment.

Once S2S/Assemble restarted, you will get this window:

This indicates you that all the "_tmp" files have replaced the old ones:

Restart Assemble/S2S one more time. You should have the new version now. I have tested this bug fix under Windows 7 and Windows Vista. It should work for Windows XP. Just let me know...

How to search for RNA motifs with Assemble 1.0.2?

2010-06-07T13:35:00.001+02:00

With Assemble 1.0.2, the "Search" lateral panel allows you to find, in your MyPDB library, the single-strands making the following topologies:

Each length of the single-strands has to be defined with the character '=', '>' or '<', followed directly by an integer value.

For example, ">4 =3 <2" means a three-way junction made with a first single-strand whose length is greater than 4 residues, a second one equal to 3 residues and a third one lower than 2 residues.

Assemble 1.0.2 is available

2010-06-07T11:29:00.010+02:00

You can get it by selecting the "Update Assemble" choice in the Help menu. The main improvements and bug fixes are the following:

- the MyPDB library is now indexed with the neo4j library. Based on this index, Assemble is able to search for three-way and four-way junctions separately. Moreover, the construction of the query has also been improved for all the categories of single-strands. Instead to search for single-strands with a precise length, we allow now the use of the characters ‘=’, ‘<’ and ‘>’ (a blog post will be published today about this topic). Don't be surprised, when Assemble 1.0.2 is launched for the first time, it takes some time to index all the 3D structures stored in the MyPDB library

- for the mouse and keyboard shortcuts, the user can now decide to use the "Ctrl" key each time the "Alt" key was necessary. This option has been added to the lateral preferences panel. For the MacOSX users, we recommend to keep the "Alt" button due to a problem with the Java implementation on Mac OS X

- for the 2D panel, the combination of "Shift key + left mouse button pressed and dragged" activates a lasso selection for the 2D panel. Depending on the selection mode activated, the lasso will select a set of "Residues", "Interactions" or "Structural Domains"

- the automatic update system has been improved. If a new version is available, the user has to confirm the update. Moreover, the update system can now add/delete files in the hidden directories ".assemble" and ".paradise". This will allow us to provide automatically new RNA motifs to the MyPDB library (for example)

- the button to restore the undo state has been recovered

Non-english characters for the comment of a new RNA motif in Assemble

2010-06-03T11:14:00.000+02:00

If you're using non-english characters for the comment of a new RNA motif (like french accents for example), you could have a problem to get its details once created. At now, we recommend you to use only english characters. This will be fixed in the next update of Assemble.

The mailing-lists of S2S and Assemble are opened

2010-06-01T13:37:00.000+02:00

I have just opened two mailing lists to allow you to post remarks and questions about S2S and Assemble:
- for Assemble, you can register here
- for S2S, you can register here

You can also use any of them to discuss about the RNA Web Services.

New screencast: How to infer a 3D model from S2S?

2010-06-01T10:58:00.004+02:00

Since the last updates (S2S 2.0.1 and Assemble 1.0.1), the connection between these two tools has been simplified and improved. When a 3D model is inferred from S2S, the 2D and 3D data are saved in the directory of the structural alignment. Then you can open Assemble and load the 3D model from this alignment.

The first S2S or Assemble application launched on a computer initiates a working session. Any tool launched further will connect automatically to this working session. Consequently, many situations can be realized:
- you can infer two different 3D models from S2S and open each one with a different Assemble application
- you can use Assemble to display the reference structure of the structural alignment (instead of PyMOL or Chimera)
- you can open and connect two S2S applications

The new screencast is available here. Don't forget that it needs (at least) S2S 2.0.1 and Assemble 1.0.1 (which are available as automatic updates from S2S 2.0 and Assemble 1.0).

If you cannot launch Assemble or S2S with Windows XP....

2010-05-25T09:47:00.001+02:00

A user reported me a problem with Windows XP (thank you Jorge for this feedback). Once Assemble installed, he got the message "problem to connect to localhost" and this froze the program.

When Assemble (or S2S) starts, the underlying paradise component initializes a network communication layer needing the port 1099. At now, this communication layer is used to exchange data between several S2S/Assemble applications and to send requests to the Web Services. If the port 1099 is already in use, this will block the launching process. Jorge found that this port was used by the WinUpdate process. Once WinUpdate killed, Assemble was able to start.

In one of the upcoming updates, i will add an option to allow a user to change the default port needed by Assemble/S2S. At now, you can use tools like TCPView to check if this port is already in use on your machine.

Assemble 1.0.1 and S2S 2.0.1 are available

2010-05-21T15:29:00.013+02:00

You can get them by selecting the "Update S2S/Assemble" choice in the Help menu of each tool. The main improvements and bug fixes are the following:

- Assemble 1.0 had problems to load a model where parts of the 2D were not present in the 3D. A workaround was to export all the residues into the 3D scene and to save the model (which is not practical if the 2D was large). Assemble 1.0.1 fixes now this problem.

Assemble 1.0: only the residues in red are present in the 3D scene.

You can see that the bases are weird.

The same model loaded by Assemble 1.0.1

- Assemble 1.0.1 provides now a button to reload the RNA Motifs library. This is useful to display the motifs created/deleted from a second Assemble application launched in parallel.

- the ability to export single residues into the 3D scene of Assemble has been restored. Once the "Residues" selection mode activated, select in the 2D panel the residues to export and click on the "Generate 3D from 2D" button.

- in S2S 2.0.1, the different selection modes are reactivated for the reference sequence in the alignment panel (Residues, Interactions, Structural Domains)

- when a tertiary structure is imported into the MyPDB library, its secondary structure is calculated during the import (in Assemble 1.0, this was done when the 3D was loaded for the first time). Consequently, once imported, this new 2D/3D can be targeted by the Search panel.

- the communication between S2S and Assemble has been improved (and will be described in more details in an upcoming blog post)

Network connection problems at the University of Strasbourg

2010-05-11T15:11:00.002+02:00

Since several days, you may experience temporary network problems by using Assemble, S2S or the RNA WebServices. The network of the University of Strasbourg is migrating to a new architecture. We apologize for the inconvenience.

Assemble 1.0 and S2S 2.0 are available

2010-04-22T15:22:00.004+02:00

The final versions of Assemble 1.0 and S2S 2.0 have been released.

We're really excited to let you discover the new features of Assemble. Among them, you will find:
- the embedded structural library named MyPDB. It provides more than 200 pre-annotated tertiary structures derived from the Protein DataBank
- the ability to search for RNA motifs in this library
- the compatibility with windows 7 and XP
- the homogeneous way to manipulate the 2D and 3D scenes

We have also improved the online documentation for Assemble. You will find now reference cards to print, a set of video tutorials and FAQs.

S2S has been less drastically updated. It is now compatible with windows 7 and XP. The inference of a new 3D model from a structural alignment has been improved. If you launch S2S and Assemble on the same computer, you can export the new 3D model directly into Assemble. Any selection made in S2S for the sequence exported will highlight its counterpart in Assemble.

You can find all the details about the improvements and the bugs fixed in the CHANGELOG files provided with each tool.

We hope that you will enjoy these new versions.

S2S featured on cover of March's issue of the Nature Methods journal

2010-03-02T09:19:00.002+01:00

I'm happy to announce that S2S has been featured on cover of March's issue of the Nature Methods journal.

The cover image shows a range of data visualizations currently used by life scientists. This issue contains a series of five commissioned reviews discussing the challenges of visualizing biological data and the visualization tools available to biologists working with genomes, alignments and phylogenies, macromolecular structures, images and systems biology data.

Update of the Web Service Annotate3D

2010-02-12T08:45:00.008+01:00

The Web Service Annotate3D has been updated. Using the "pdbid" option, it can download data directly from the Protein Data Bank. To do so, the "data" option of the Java client has to be equal to "-". The Java client has been updated, you can download it here.

Example of usage:
java -jar client.jar ws=annotate3d data=- pdbid=1HR2

You can also use command-line tools like wget (Linux) or curl (MacOSX):

wget -q -O - "http://paradise-ibmc.u-strasbg.fr/webservices/annotate3d?pdbid=1HR2"
curl -s -o - "http://paradise-ibmc.u-strasbg.fr/webservices/annotate3d?pdbid=1HR2"

or your favorite Web Browser by typing web adresses like:
http://paradise-ibmc.u-strasbg.fr/webservices/annotate3d?pdbid=1HR2

Update of the RNA Web Services

2010-02-02T14:06:00.006+01:00

Today, i have improved the way to call the RNA Web Services. In general, biologists know what they want to do (predict a secondary structure, annotate a 3D structure, produce an alignment,...). But they don't know which algorithm can help them. And, frequently, they don't care about the name of an algorithm, they simply want a result to pursue their reflexion. Consequently, the way to call the RNA Web Services using "algorithm_name/action_name" was not the most intuitive one. I renamed them and i only kept the "action_name" (predict2d, plot2d,...). If several algorithms are/will be available for a given Web Service, a "tool" option allows the user to define one. If it is omitted, a default algorithm will do the job.

To reflect these changes, I have updated the online documentation and the Java tool. I have also updated the posts on this blog providing examples of usage for these Web Services. By the way, i have registered them into the biocatalogue repository.

How to load the tRNASec embedded in the crystal structure of the human SepSecS-tRNASec complex?

2010-01-31T15:20:00.009+01:00

Today, i have observed that S2S or Assemble cannot open the PDB file 3HL2 describing the crystal structure of the human SepSecS-tRNASec (Palioura et al, Science, 2009). This is due to the RNAVIEW algorithm which is not able to annotate the 3D structure. Two conformers (A and B) of the tRNASec are described in this file. To fix this problem, you have to export one of them in a new file. To do it for the conformer A, just type in a UNIX terminal:

grep -E "^ATOM.{12}A.+$" 3HL2.pdb | sed -E 's/(^ATOM.{12})A(.+)$/\1 \2/' > tRNASec_A.pdb (for MacOSX)
grep -E "^ATOM.{12}A.+$" 3HL2.pdb | sed -r 's/(^ATOM.{12})A(.+)$/\1 \2/' > tRNASec_A.pdb (for Linux)

Now you should be able to load this conformer by opening the tRNASec_A.pdb file with S2S/Assemble.

What is PARADISE?

2010-01-28T08:12:00.028+01:00

I observed from discussions i had with S2S or Assemble users, or by reading publications citing these tools, that there is a confusion with the name PARADISE. I will try to explain it with this post. S2S and Assemble are two graphical tools constructed above a common layer. This layer is mainly composed of two parts:
- the implementation of the biological concepts manipulated through the graphical tools (tertiary and secondary structures, structural alignments, atoms, helices, nucleotides,...)
- the implementation of a communication protocol to allow the graphical tools to interact with each other (this aspect will be improved with the next releases) and with external elements (like the RNA Web Services).

This layer is PARADISE. Why did I gave an explicit name to this part of my framework? Because it is possible to use PARADISE without the graphical tools. If you're looking in the "lib" folder of your S2S or Assemble, you will see a paradise.jar file. Inside this "lib" folder, you can type in a terminal:


java -jar paradise.jar

to start a scripting console. This console allows you to write scripts using all the biological objects and the communication protocols developed so far. These scripts have to be written with the Groovy language (also used to write your own S2S/Assemble plugins). I'm using this console to do some batch processing.

Now you should understand better what PARADISE is. It's not S2S, not Assemble, it's a part of them.

PS: for those interested in the technical details, the communication protocol is based on a multi-agent framework implemented using the JADE library

Registration for RNA2010 is now open!

2010-01-28T07:49:00.000+01:00

The registration for RNA2010 is open. The 15th annual meeting of the RNA Society will take place on the University of Washington campus in Seattle, WA from June 22nd-26th, 2010. The deadline for submission of abstracts and registration is Monday, March 15th, 2010.

Using the new RNA Web Services with Mathematica

2010-01-23T15:18:00.021+01:00

To send a request to the RNA Web Services, you need to construct an HTTP POST message. Unfortunately, the Import function of Mathematica always uses the GET method. As highlighted recently on the Mathematica blog, the HTTP POST can be performed with Java which is "JLinked" to Mathematica. To predict a secondary structure for an RNA sequence stored in a FASTA file, just do the following:

<< JLink`
client = JavaNew["org.apache.commons.httpclient.HttpClient"];
data = Import["my_rna.fasta", "String"];
method = JavaNew[
"org.apache.commons.httpclient.methods.PostMethod",
"http://paradise-ibmc.u-strasbg.fr/webservices/predict2d?"
];
method@addParameter["data", data];
client@executeMethod[method];
ImportString[method@getResponseBodyAsString[]]

Batch processing with the new RNA Web Services

2010-01-21T21:56:00.017+01:00

Imagine that you have several PDB files for which you would like to compute a secondary structure, you can type in a UNIX terminal (as a single-line command):

for f in *.pdb; do java -jar client.jar ws=annotate3d data=$f > $f.ps; done

Each secondary structure will be saved in a Postscript file whose name is identical to the corresponding PDB file with a ".ps" suffix.

To predict an RNA secondary structure for several sequences stored in FASTA files, type (as a single-line command):

for f in *.fasta; do java -jar client.jar ws=predict2d data=$f > $f.bseq; done

To do large batch processing, I suggest you to increase the memory allocated to the java client. For example, to allocate 500M:

java -jar -Xms500M -Xmx500M client.jar ws=webservice_name ....

The RNAVIEW Web Service has been fixed

2010-01-20T18:28:00.009+01:00

Perhaps have you observed that RNA 3D annotations done recently with Assemble or S2S produced the following message for any PDB opened:

"No base-pairs found, no 2D structure plot generated (may be a single-strand??)"

This problem is more or less related to the release of the new version of the Web Services (as announced in the previous post). This should be fixed now. Don't hesitate to contact me if you observe other strange behaviours.

The new RNA Web Services are available

2010-01-18T08:24:00.013+01:00

During the last few weeks, I have implemented a new version of the Web Services wrapping RNA algorithms. In contrast to the previous version, they can also be used without S2S and Assemble. Since they're now following the REST way, they can be invoked from a simple web browser.

I also provide a really simple java client to send data that cannot be enclosed easily in a Web address. This client supports the standard input/ouput and consequently can be piped with other unix commands (take a look at the online documentation for samples).

These new Web Services will be used by the upcoming releases of S2S and Assemble. Soon, new RNA algorithms and new options to configure them will be available.

Don't hesitate to test them and to send me feedback.

One day perhaps...

2009-11-20T09:57:00.002+01:00

...but at now it's only a fake ;)

How to download all the RNA basepair exemplars at once from the FR3D website?

2009-11-03T07:33:00.015+01:00

The authors of the FR3D project provide a set of exemplars for each of the 12 RNA basepair families. To get all of them as PDB files, you can type in a terminal:

- for Linux:

wget -O - "http://rna.bgsu.edu/FR3D/basepairs/PDBExemplars/" | grep ".pdb" | sed -r 's/.+<a href="(.+)">.+/\1/' | xargs -I % wget -O % "http://rna.bgsu.edu/FR3D/basepairs/PDBExemplars/%"

- for MacOSX:

curl -o - "http://rna.bgsu.edu/FR3D/basepairs/PDBExemplars/" | grep ".pdb" | sed -E 's/.+<a href="(.+)">.+/\1/' | xargs -I % curl -o % "http://rna.bgsu.edu/FR3D/basepairs/PDBExemplars/%"

In both cases, it is a single command line.

Assemble and Mac OSX 10.6 "Snow Leopard"

2009-09-18T15:17:00.013+02:00

If you have tested Assemble with Mac OSX 10.6 "Snow Leopard", you have perhaps observed that Assemble is not able to start. This is due to the fact that Snow Leopard is using the 1.6 version of Java instead of the 1.5 one. If you check the web, you will discover that they are many issues with Java on the new Mac OSX 10.6 "Snow Leopard".

To fix this problem, just download and unzip this file. Move the jogl.jar and gluegen-rt.jar files into the lib directory of Assemble. Move all the other files with the jnilib suffix into the native/mac directory. Once restarted, Assemble should now work with the new cat from Apple ;)

The Release Candidate 3 of Assemble 1.0 and S2S 2.0 are available

2009-08-05T10:31:00.005+02:00

I'm happy to announce the availability of the Release Candidate 3 (RC3) of S2S 2.0 and Assemble 1.0. This new release is mainly focused on the ability of S2S and Assemble to interact with RNA algorithms exposed as Web Services.

Due to this important modification, you cannot get these new versions with the built-in update system of the previous releases. You have to download and install them in a new directory on your computer. If you have subscribed to the newsletter, you got in your mailbox the direct links to download them. If not, you can get them from the download pages of S2S or Assemble.

Follow these links to see the bugs fixed and the improvements:
- for S2S: http://ftp.bioinformatics.org/pub/s2s/2.0_RC3/CHANGELOG
- for Assemble: http://ftp.bioinformatics.org/pub/assemble/1.0_RC3/CHANGELOG

New features to edit an RNA secondary structure

2009-07-31T12:10:00.019+02:00

These new features will be available with the Release Candidate 3 of S2S and Assemble (you will be able to download them in few days, perhaps this week-end ;)):

- when an helix is removed in the 2D panel, the secondary interactions can be kept as tertiary ones
- a new helix can be created at any position, even if it overlaps a previous one. The 2D panel will remove the overlapping helices, and transform their secondary interactions as tertiary ones (a consequence of the previous feature). Then, the new helix and its secondary interactions are created. If a new secondary interaction overlaps a tertiary one, the secondary interaction substitutes it.

So, for example, you can imagine this situation with Assemble (you will have the same result with S2S):

I'm selecting four residues to create a new helix:

By default, the orientation of a new helix is vertical (that's not the best, I will improve that one day):

But if you ask Assemble to reorganize automatically the helices, you get:

To convince you that new secondary interactions substitute overlapping ones, I'm modifying an AU to a transWatsonCrick-SugarEdge and I'm adding a new tertiary interaction (AA cisHoogsteen-SugarEdge):

Now I would like to extend my last helix, with two additional base-pairs at one end (with the AA cisHoogsteen-SugarEdge) and I would like to exclude the last UG base-pair at the other end. Consequently, my four-bases selection will be the following one:

Once the helix created, a reorganization of helices produces me this result:

You can see that I have an extended helix with the base-pairs as I defined them in the previous steps. The last UG is now a tertiary interaction.

Update of the Assemble documentation

2009-07-24T14:47:00.002+02:00

The page describing how to use Assemble has been improved. I have followed the same structure than for S2S.

Forget the installation of RNA algorithms, now it's only Web Services!!!

2009-07-23T09:09:00.010+02:00

Due to many problems with the installation and usage of RNA algorithms, I have decided that S2S and Assemble will interact only with RNA algorithms exposed as Web Services.

As highlighted with the previous post, some are already available. I have released my own Web Services for the remaining ones. They have been developed with the Apache Axis library and are hosted by the servlet runner Apache Tomcat 6. You can find all the details on this webpage. At now, these Web Services are working with the default parameters of the original algorithms and you cannot modify them. But I will improve them ASAP.

The direct consequences of this new orientation are:
- it's not anymore necessary to install the RNA algorithms on your machine,
- it's not anymore necessary to start S2S or Assemble from their own directory (as described here)

I hope that this will improve the usability of S2S and Assemble.

S2S and Assemble can now consume Web Services

2009-07-13T14:23:00.013+02:00

For the Release Candidate 3 of S2S and Assemble, I have modified the underlying engine (a.k.a. PARADISE engine) to be able to consume Web Services (for those interested by the technical details, i have used the JAX-WS implementation to do the job). The PARADISE engine is now able to use the vrnaplot and vrnafold Web Services provided by the EBI. Vrnaplot and vrnafold are ports of the Vienna RNA package programs provided by the EMBOSS suite.

These Web Services will be a good alternative if you have problems to run a local algorithm. The final goal is to propose this alternative for any algorithm used by S2S or Assemble.

I have also decided to remove the "Disconnected mode" proposed at the startup. The graphical tool has to connect a local or a remote platform. The capability to do a task will depend on the ability of the connected platform to find at least one service linked to a working algorithm (locally or as a Web Service).

The Release Candidate 2 of Assemble 1.0 and S2S 2.0 are available

2009-07-06T11:06:00.003+02:00

I'm happy to announce the availability of the Release Candidate 2 (RC2) of S2S 2.0 and Assemble 1.0.

If you didn't get it until now, download the RC1 at these adresses:

- for S2S: http://www.bioinformatics.org/S2S/download.php
- for Assemble: http://www.bioinformatics.org/assemble/download.php

Install it and update it to the RC2 version (menu choice Help -> Update).

Follow these links to see the key features of these new versions:

- for S2S: http://ftp.bioinformatics.org/pub/s2s/2.0_RC2/CHANGELOG
- for Assemble: http://ftp.bioinformatics.org/pub/assemble/1.0_RC2/CHANGELOG

PLEASE NOTE AN IMPORTANT POINT:
------------------------------------------------------
To be able to reload the data saved with the RC1 versions, they need to be saved using the textual format (and not the binary one). If this is not the case, resave them before to update. I remember you that the textual format allows you to be compatible between the different versions of S2S and Assemble. I suggest you to always keep a textual version of them. The binary format increases the speed of loading/saving data. But they're linked to the version which produced them.

I hope that you will enjoy these new versions.

Load a secondary structure described with the bracket notation

2009-06-24T15:27:00.005+02:00

With their new release, S2S and Assemble will be able to load a secondary structure stored in a fasta file and described using the bracket notation.

The fasta file has to follow these rules:


>Secondary Structure
.......((((.....(((((((.((...((.(((........))).)).....)).))))...)))..((((..(((..(((....((((...(((((...))))).....))))..)))..))).......))))........)))).
>RNA test
UAACAAUCUGCUGAAAGGUACCGUCGGAGGGAGCUUUGUUGCCAGCGCCAGAAACGCCGGUUUAACCAGCGCCGAAGUGAGCGCAGUGAUUAAAGCCAUGCAGUGGCAAAUGGAUUUCCGCAAACUGAAAAAAGGCGAUGAAUUUGCGGU

The bracket notation has to be preceded by the ">Secondary Structure" line.

S2S and Assemble are now available to anyone

2009-06-15T11:59:00.003+02:00

Everything is in the title. It's here for S2S and here for Assemble. Enjoy!!

How to launch S2S and Assemble efficiently?

2009-06-11T08:42:00.031+02:00

Some users have encountered a well-know problem (at least for me ;)) during the startup of S2S or Assemble. When they wanted to open a secondary structure (stored in a CT or BPSEQ file), or to predict one (by opening a FASTA file), they got an error like:

The RNAPlot-1.7 service is not able to process your request efficiently
Error
Message:
fr.unistra.ibmc.paradise.goloka.core.analysis.AnalysisException: fr.unistra.ibmc.paradise.goloka.core.analysis.AnalysisException: Unable to find the result SVG file
[...]

At now, to be able to find the RNAPlot output, S2S and Assemble have to be launched from their own directory. Two ways to do that:

by double-clicking on the assemble.jar or s2s.jar file
by typing in a terminal: "java -jar assemble.jar" or "java -jar s2s.jar"

For the second option, you could also do something like:

java -jar my_applications/S2S/s2s.jar

And then you will get the error.

So, how it is possible to launch S2S or Assemble from any directory? If we suppose that the location of S2S is /home/fjossinet/my_applications/S2S, you can follow these steps:

put the following lines in a script named s2s.sh (for example):

#!/bin/bash cd /home/fjossinet/my_applications/S2S java -jar s2s.jar
set the rights to execute the file for the owner of this file (you!!): chmod u+x s2s.sh
move this script in a directory registered in your PATH
type s2s.sh from any directory and open a secondary structure. It should work.

If your S2S directory is located under your home directory, this version is more elegant:

#!/bin/bash
cd $HOME/my_applications/S2S
java -jar s2s.jar

The advantage of this approach is that you can also easily modify the memory allocated to your java environment. Reopen your script and modify it like that:

#!/bin/bash
cd $HOME/my_applications/S2S
java -Xms500M -Xmx500M -jar s2s.jar

to allocate 500M for example.

With the next release, S2S and Assemble will warn you if you don't start them from their own directory. I will also see how to tweak RNAplot to remove this technical constraint.

Using SAXS data with Assemble

2009-06-08T11:20:00.012+02:00

By using the Situs package, you can use the ability of Assemble to load density maps to display your SAXS data.
If you have your data stored in a sample.pdb file, do the following in a terminal:

pdb2vol sample.pdb sample.situs

then:

map2map sample.situs sample.mrc

For my test, I have converted my file to the MRC format (choices n8 or 9 with the 2.4.3 version of Situs, both are working).

Open Assemble, load a 3D model and then load your sample.mrc file. If necessary, modify the parameters provided by the Map panel to see your map.

The Release Candidate 1 of Assemble has a bug that will be fixed with the next update. If you cannot see your MRC file with the open dialog, type its name along with its absolute path directly in the open dialog.

For the next release, I have also updated the X-PLOR parser to be able to load the X-PLOR file created by Situs (choice n7 with the 2.4.3 version of Situs). At now, you will get an error.

Manipulating RNA secondary structures predicted by mfold with S2S or Assemble

2009-05-29T13:23:00.004+02:00

S2S and Assemble delegate the 2D prediction to RNAFold from the Vienna RNA Package. I'm planning to add mfold to the calculation layer. For the moment, you can use the mfold server. S2S and Assemble are able to read any CT format provided by this server: .ct file, RnaViz ct and Mac ct.

The Undo History Panel of Assemble

2009-05-29T08:39:00.003+02:00

For the Release Candidate 2 of Assemble, I have improved its Undo System. Assemble provides now an "Undo History Panel". Each time the user clicks on the "Accept current 3D model" button (in the 3D scene toolbar), the current state of the 3D model is saved with a user-defined name and added to the "Undo History Panel".

By double-clicking on an entry in this panel, the user recovers the corresponding state. The new Undo System can keep track of the 3D coordinates but also of the residues created or deleted between two consecutive states. It also saves the "cut/bind molecular chains" actions.

When a 3D model is saved as an Assemble model (using the textual or binary format), the "Undo History Panel" is reinitialized. The saving process synchronizes the underlying model with the 3D view. Consequently, some saved states could be incompatible with this new underlying model. The Undo History is conserved if this 3D model is exported as a PDB file.

S2S poster

2009-05-28T09:51:00.004+02:00

I did a poster about S2S for a meeting organized in Strasbourg (Second European Meeting on Yeast and Filamentous Fungi). You can get a PDF version here.

The Protein Data Bank is using Twitter

2009-05-13T21:46:00.003+02:00

To follow them, go there.

A code snippet to rename a set of pdb files

2009-05-04T13:39:00.016+02:00

If you have recovered a set of PDB files from the NDB or PDB databases and if they have "pdb1" or "pdb2" as suffix, do the following in a terminal to have "pdb" instead:

for f in $(ls *.pdb[12]); do mv $f ${f:0:${#f}-1}; done

The RNA algorithms provided with Assemble and S2S

2009-05-04T10:51:00.015+02:00

Assemble and S2S delegate the important tasks to RNA algorithms hosted on a calculation layer. I'm naming this calculation layer a "PARADISE platform". It is based on an "RNA engine" named PARADISE which gathers two main parts:

a vocabulary: the RNA concepts displayed and manipulated (helices, single-strands, tertiary and secondary structures, structural alignments,...)

a communication layer to interact with the RNA algorithms producing data based on this vocabulary. This layer is itself constructed on a multi-agent system using the JADE library

As I will describe it later on this blog, this "RNA engine" can be used to produce new tools. Assemble and S2S can be seen as the combination of this "RNA engine" with a dedicated graphical interface.

To get the full functionalities of Assemble and S2S, the user has to connect them to such a PARADISE platform. If not, they can work in a "disconnected" mode with reduced abilities.

With the previous versions (before the Release Candidate 1), the user had to start such a platform in a first terminal, and then the graphical interface in a second one. I changed this behaviour to reduce the complexity of the launching process. Now the user starts the graphical tool only. If "localhost" is chosen in the splash screen, then a platform is transparently started on the machine of the user, and the graphical interface tries to connect it. The user has also the choice to contact a remote platform by registering its IP address in the splash screen. In this situation, an Assemble or an S2S has to be started with the "localhost" option on the remote machine. This can be seen as a kind of "peer-to-peer" system. This could be practical for using the graphical tools on a Windows machine connected to the RNA algorithms of a UNIX one.

To have a working PARADISE platform, the RNA algorithms hosted on it have to be compiled. At now, a platform provides RNAVIEW to annotate RNA tertiary structures and the Vienna RNA package for the 2D prediction and drawing. There are also Java algorithms developed in our lab (3D prediction and 3D refinement) that don't need to be compiled. Compilation of source code can be a tricky task. I wanted to avoid this difficulties to the users by providing a public PARADISE platform. One week before the availability of the Release Candidate 1, i observed that the communication with this public platform was too slow, mainly due to our firewall.

Consequently, i decided to provide the RNA algorithms already compiled for three different architectures: MacOSX on PowerG5 (ppc) and Intel (i386) processors and Linux on Intel (i386) processors. They're located in the lib/algorithms subdirectory. If you're using an other architecture, S2S or Assemble will warn you that it is not supported. Then you can go into the lib subdirectory and start the algorithms_installation.sh script. If everything goes well, you should find your binaries in a new directory named "your_os/your_processor" in "lib/algorithms".

With the Release Candidate 1 of S2S and Assemble, I have provided the binaries for the three architectures in the tar.gz file. To reduce the size of this file, I will implement a way to automatically download the binaries needed during the first launch.

Books added to my reading list

2009-05-03T15:54:00.004+02:00

Release Candidate 1 for S2S 2.0 and Assemble 1.0

2009-05-01T22:13:00.009+02:00

Yesterday evening, I sent to the "official" beta-testers of S2S and Assemble the instructions to get these new versions. It was the last day of April, but i respected the deadline announced recently ;) I hope that they will enjoy the work done...
(05-02-2009 update: i discovered today that the Release Candidate 1 of Windows 7 was also available on April 30th. I'm not a big fan of this OS, but the coincidence is funny)

The Stockholm format

2009-04-29T11:23:00.008+02:00

S2S is now able to parse RNA alignments stored with the Stockholm format. Consequently, I can open the RNA alignments and secondary structures of the Rfam database. At now, the reference secondary structure is the first sequence found in the file. Here is for example the display of the Lysine riboswitch (entry RF00168). The secondary structure displayed is for the sequence CP000002/2878377-2878198:

The S2S documentation is online

2009-04-29T11:18:00.003+02:00

At last.... You can find it here. To me, it is more understandable than the Assemble documentation. I will improve it also.

News Panel for S2S and Assemble

2009-04-21T14:13:00.000+02:00

This is the last feature I wanted to implement this week before to come back to more "bioinformatics" improvements. I have added a little panel in the GUI of S2S and Assemble to display the news I publish about these tools. More precisely, it recovers the posts on this blog. The RSS channel is checked each 30 minutes. Double-click on an entry to read the post in your browser:

At now, I didn't filter the posts, so all of them are recovered.

Automatic Update System for S2S and Assemble (Part2)

2009-04-21T11:20:00.000+02:00

This morning, I have improved a little bit this update system. When S2S or Assemble start, you have this kind of splash screen:

If the user chooses the "disconnected" mode or wants to be connected to his local platform, then no check for an update is done automatically (in general, you're up to date with yourself ;)). But this update can be asked for, once the tool started, using the "Update" choice in a menu.

If the user chooses to connect his graphical tool to an external platform, a check for an update is done automatically. If a new version is available, the tool is updated and quit. The user gets the new version once the tool restarted.

Why did I implement this behaviour? If you know a little bit the architecture behind S2S and Assemble, you remember that these tools delegate some work to RNA algorithms hosted on a remote calculation layer (a.k.a "a platform"). The communication is based on a Multi-Agent System and is implemented using the JADE library. You have two ways to exchange the data: the textual and the binary formats. I'm using the second one. Due to the complexity of the data exchanged, this improves the speed of communication. But the server and the clients need to have the same version of the library to be able to serialize/unserialize the Java objects exchanged. Consequently, the default behavior of a graphical tool connected to an external platform is to get the most recent version available, before to initiate this connection. Then, it's up to the administrator of a platform to update it frequently. Otherwise, the user of an updated tool will see a message like: "The platform is not up to date with your tool. Contact its administrator to update it."

Automatic Update System for S2S and Assemble

2009-04-20T10:52:00.000+02:00

Today, I have implemented a very simple system to download automatically a new release. When S2S and Assemble start, they are parsing a remote XML file to check if a new version is available. At now, it's "quick and dirty" (not so dirty I hope ;) ).

To my own experience, biologists don't like to deal with concurrent versions system and/or continuous integration engines. Consequently, I will use this system to diffuse the important evolutions of my tools.

3D RNA Modeling Workshop

2009-04-20T07:51:00.000+02:00

This week-end, i received this event to publish on the RNASociety website:

This workshop is related to the last publication of Dr Russ Altman's group: Coarse-grained modeling of large RNA molecules with knowledge-based potentials and structural filters.

Reference Manager Overview

2009-04-19T10:57:00.000+02:00

By following Neil Saunders on Twitter, I discovered this morning this matrix chart comparing the features of the different reference managers available:

I didn't find BookEnds for OSX. But I discovered Mendeley. I should try it.

Use Assemble and S2S to generate interactively subsequences and substructures (Part1)

2009-04-16T08:44:00.001+02:00

In the File menu of Assemble, you will find several options to export your data. You can export the 3D model in a PDB file, the 2D model in a CT file or the molecules of your models in a FASTA file. You will also find a "Selection..." sub-menu to export the selection you made in the 2D or 3D models.

The ability to export a selection should be interesting for several reasons:

export the sub-domain of a 3D structure (like an RNA motif) to study it without the need to keep the full structure in memory

export the sub-domain of a 2D structure to use it as a new start for the construction of a 3D model

export a sub-sequence in a FASTA file to produce a new 2D fold or to find homologuous sequences in databases

To illustrate these features, I decided to load the Crystal structure of the bacterial ribosome from Escherichia coli at 3.5 A resolution. At now, loading this huge structure with Assemble takes a while. I have to improve the speed of the software. Once loaded, annotated by RNAVIEW and displayed in the 3D scene, I get this result:

As you can see, the 2D structure produced by the RNAVIEW algorithm is a flattened version of the 3D. This 2D is difficult to manage in this state. Consequently, I'm asking Assemble for a reorganization of the helices. This option calls the RNAPlot algorithm from the Vienna package and produces a non-overlapping 2D drawing. The result is the following:

Now, I can more easily select a structural domain (in red in the 2D model and in yellow in the 3D model) and export it as a new PDB file

When I load the PDB file just created, I get the following result which is easier to manipulate (for humans and for computers ;))

With the ability to export a 2D selection in a new CT file, you can also create new kinds of 2D. The CT file has been designed to store the informations for a single molecule. Consequently, if i select and export discontiguous regions from my original 2D:

I get this result if i reload the CT file created:

As you can see, due to the CT file format limitations, i have lost the tertiary interactions and all the base-pairs are described as classical cis-WatsonCrick-WatsonCrick interactions.
In the Part 2 of this post, I will illustrate the selection's export in a FASTA file with S2S.

Editing an RNA Secondary Structure with Assemble (Part 2)

2009-04-14T08:28:00.001+02:00

The idea of this second part has been suggested by an Assemble user (thank you Isabelle ;)). The problem is the following: you want to construct a 3D model containing two (or more) RNA molecules. These molecules can make some intermolecular interactions and helices. You can imagine for example something like a kissing-complex.

At now, your options are the following:

find a resolved tertiary structure and edit the 2D structure calculated and displayed by Assemble. But you need to find a 3D structure with the RNA sequences you want (the edition of an RNA sequence is not available in Assemble and will be discussed in an other post)
construct your 2D yourself.

For the second option, you have to provide a CT or BPSEQ file describing a first draft of 2D. But CT and BPSEQ formats describe single-molecular 2D. You can also decide to use a FASTA file describing your sequences and let Assemble predict a 2D. But, at now, Assemble delegates the 2D prediction to RNAFold which work only on a single molecule. Consequently, I added to Assemble the ability to extend the current 2D model with new RNA molecules.

In the 1.0 version of Assemble, you will find a menu named "2D model" containing the "Import" choice. If you decide to import an RNA molecule described in a FASTA file, Assemble will calculate a 2D for this new molecule and add it to the current 2D. This menu proposes you also to import a pre-calculated 2D described in a CT or BPSEQ file. Once the new 2D imported you can edit it and add inter-molecular interactions as described in the previous post.

To construct the 3D model based on the kissing complex cited above, i created the following FASTA file:

>HIV-1 subtype A
CUUGCUGAGGUGCACAGCAAG

When i opened it with Assemble, it predicted me the following 2D:

Then i added the same molecule using the "import RNA molecule" function:

Logically, Assemble predicted me the same folding:

By selecting its four ends, I created dynamically the inter-molecular helix:

Then, I generated the first draft of 3D containing the 3 helices and their connected single-strands (the inter-molecular helix is on the right):

The construction of the 3D model with the tools provided by Assemble will be the topic of upcoming posts.

Researchers develop new way to see single RNA molecules inside living cells

2009-04-08T09:29:00.000+02:00

A really interesting article about a new type of probe developed by Santangelo and colleagues. It allows to visualize single ribonucleic acid (RNA) molecules within live cells. The original article has been published online in the journal Nature Methods on April 6.

Editing an RNA Secondary Structure with Assemble (Part 1)

2009-04-02T10:12:00.000+02:00

The construction of an RNA 3D model starts with the definition of an RNA secondary structure (which is, at least energetically, the main component of an RNA architecture). Using Assemble, you can:

open a file describing such a structure (described in a CT or BPSEQ file),
open a resolved tertiary structure described in a PDB file (and Assemble will call the RNAVIEW algorithm to annotate it with a secondary structure),
open an RNA sequence described in a FASTA file (and Assemble will call RNAFold from the Vienna Package to predict a secondary structure),

In general, this secondary structure is produced by an RNA algorithm. Biologists use this result as a first draft. Then, they have to adapt it to fit their experimental data. The structural edition in the 2D panel of Assemble was already available in the last developement releases. This week I have improved it.

To give an example, I decided to use the third option and to predict and edit the secondary structure of a non-coding RNA candidate. I went to the RNAdb database and I downloaded the entry FT3100 from the FANTOM3 dataset. Once opened with Assemble, the RNAFold algorithm proposes me this fold:

You can select one or several helices and delete them:

Then you can select four residues to define a new helix. Each couple of residues has to be located in the same single-strand and their distance has to be greater or equal to 2 (this will produce an helix of size >= 2):

The design of a pseudoknot is possible:

You can change the type of base-base interaction:

Or add new tertiary interactions:

The tertiary interactions and the type of the base-base interactions will be important when you will refine your 3D model. The refinement algorithm will improve your models according to the base-base interactions defined in the secondary structure.
Last step, select in your secondary structure one or several structural domains (helix or single-strand) to produce their 3D fold:

At any moment, you can delete a 3D fold, edit the secondary structure and generate a new one.
This improved RNA 2D edition will also be available for S2S 2.0.
In an upcoming post, I will show you how to construct a secondary structure with data coming from different files.

The countdown has started

2009-03-27T14:13:00.001+01:00

Ok... I have announced on the S2S and Assemble websites that I will release their new versions for the second half of April 2009. Less than one month to find and fix as many bugs as possible ;)

Introduction

2009-03-23T13:07:00.001+01:00

To start this blog, several informations about me. I'm an assistant professor in a french university since 2002. I'm also a molecular biologist doing software development since 2001 (since the end of my PhD more precisely). So far, I developed Java tools for structural biology (check the S2S and Assemble websites for details). Beside biology, i really like to discover and test new programming languages: python, ruby, groovy, awk, Cocoa (my favorite development environment is OSX), Mathematica (a little bit),.... I will post on this blog my ideas/thoughts and the progresses of my research and development projects. Stay in touch ;)