What's New in Version 4.71?

Michael A. Kappler
Daylight

 

Abstract: This presentation describes changes and features specific to version 4.71 of Daylight Chemical Information Systems software. Version 4.71 includes the Oracle Cartridge on Solaris, updates to CLOGP, JavaGrins, Thor, DayCGI, and numerous bug fixes.


Getting Started

Is version 4.71 available?

Yes. The Daylight Krewe has been working hard on this release for some time, and we know you're anxious. We're thankful for your patience.

How do I get it?

  • Arrange to obtain a Daylight distribution CD at this meeting.
  • Download the distribution from http://www.daylight.com/cgi-bin/download.cgi.
  • Email us at info@daylight.com.
  • Phone us at 1-949-367-9990.
  • What about licensing?

    Contact Yosi Taitz (CEO) at this meeting, or at the email address or phone number listed above.


    Announcements

    Chemistry Cartridge for the Oracle Database Server Environment

  • In a press release on July 14, 2000, Daylight officially introduced the Daylight Chemistry Cartridge, DayCartTM.

    DayCart is the result of several distinct collaborative development projects within Oracle, Novartis, Ontogen and Daylight. The Oracle Data Cartridge is a bundled set of tools which extends Oracle clients and servers with new capabilities.

    The history of DayCart goes back to Samuel DeFazio's MUG '98 talk, progressing through Dave Weininger's EMUG '98 talk and MUG '99 talk, and Jack Delany's EMUG '99 talk, and gaining much attention with Jack's MUG '00 presentation and beta version release. Introduction of DayCart represents a new direction for Daylight in handling corporate chemical infornmation. DayCart is current supported on SUN Solaris 2.7 systems only. Talk to Jack about other upcoming supported systems.

    The current status of DayCart and plans for upcoming supported platforms will be presented by Jack Tuesday afternoon.

    Availibility and licensing requirements will be presented by Yosi Tuesday afternoon.

  • CLOGP Without Missing Fragments

  • CLOGP 4.0 is a significant upgrade with the new capability of estimating values for unknown fragments. For more informaion, see Al Leo's presented at MUG '00.

    CLOGP 4.0 from BioByte has been fully integrated into the PCModels and PCFilters products. As a result, CLOGP is able to generate high quality calculations for many more compounds, and useful calculations for all compounds in many typical organic datasets.

  • JavaGRINS SMARTS, Simultaneous Use, Windows Native, and Speed Optimization

    Daylight's graphical molecular Editor, JavaGRINS, continues to be popular because it allows chemists to draw structures instead of typing SMILES. There are several new JaveGRINS features.

  • Substructure search query are graphically generated. The SMARTS module is completely graphical and requires no knowledge of SMARTS syntax.

  • Configuration now includes both simultaneous user licensing as well as IP-based licensing. Simultaneous user licensing configuration makes it possible to use JavaGRINS in DHCP networks.

  • JavaGRINS is offered as a native Microsoft Windows client application. The client application allows copy and paste to a SMILES or SMARTS text window.

  • Integration into a web-page is now easier than ever! Just add three lines of code to a web page.

  • JavaGRINS has been optimized for speed.

  • The DayUtils server now uses native sockets for communication (RMI has been abandoned), and responds faster than before.
  • Bigger Thor Databases

  • The Thor database size limit is now 16GB on IRIX 6.4 and higher and Solaris 2.6 and higher.

    The thorserver simulates 34-bit file access, yet the database format remains 32-bit. Therefore, the 34-bit thorserver is compatible with existing databases. The transition to larger databases is seamless. To exceed the old limit of 2GB, just start using the new product. For more information, see Michael A. Kappler's presentation titled A "Bit" Larger Database given at MUG `00.

  • DayCGI - smi2gif layout, stereo isomers, Hydrogen coordinates

  • The smi2gif program has been enhanced so that depictions can reflect stored 2D coordinates.

    When passed a TDT string (supported since 4.61) containing 2D coordinates and the "#t" option, all hex-encoded as usual, smi2gif will use supplied coordinates instead of generating a layout with the dt_calcxy() routine.

  • Stereo-isomeric SMILES are handled.

  • hydrogen coordinates will be used if they are present in correct SMILES order (none, some, or all is ok).

  • Big Fixes

    SMILES Changes

    There are several changes to the SMILES syntax.

  • Canonical SMILES will now be output with the "-" bond in cases where there is a single bond between two aromatic atoms. The prototypical case is fluorene; its canonical SMILES in 4.71 is: C1c2ccccc2-c3ccccc13. This fixes a problem with ambiguous kekule assignments which cause oscillating canonical SMILES.

  • Four-valent neutral nitrogen can no longer be aromatic. Prior to version 4.71, [Fe]n1ccccc1 was allowed as input. This was in part due to the idea of a dative bond, which was (poorly) supported in earlier toolkit versions. Unfortunately, these ambiguous valences cause other problems (such as the failure in 4.6x of [Fe]n1cccc1 to parse). In 4.71, the semantics and valences for Nitrogen have been made consistant. [Fe]N1=CC=CC=C1 is the proper input SMILES for Iron-pyridine in 4.71.

  • Nitrogen anion is now possibly aromatic. The prototypical pyrole anion (c1cc[n-]c1) is correctly handled. This fixes several reported bugs.

  • Unusual valence aromatics. Several classes of molecules would be written out as aromatic by the toolkit in 4.6x but could not be read back in. These were the result of a mismatch in the dt_smilin() and dt_cansmiles() rules for aromaticity. These have been unified.

  • Uncorrelated EZ stereochemistry. In some cases, canonical SMILES didn't get these right. They are fixed.

  • Oscillating SMILES. Most have been fixed. See 'Known Bugs' in the release notes for one known exception.

  • Relative stereochemistry (Fischer). This has been added. Basically, the toolkit will now preserve multiple relative chiralities on ring systems.

  • Extended aromaticity perception. The toolkit will now perform a combinatorial analysis of ring systems, when necessary, to identify extended aromatic systems. This is not perfect (see known bugs), but is better than 4.6x at finding extended aromatic systems.
  • Rubicon Changes

    The rules for invention of conformations has been expanded for a wider range of support for large and small molecules. Missing parameters are becoming less of an issue.

  • Enhanced hybridization recognition has improved the geometry of isonitriles, triazoles, carbon anions, dioxides, and diazo compounds. Also, planarity of aromatic rings with more than six atoms and embedded cyclobutadienes is enforced.

  • Rubicon now rates second in a 3D conformation generation benchmark. Using the test set described in: Sadowski, Gasteiger, and Klebe; J. Chem. Inf. Comput. Sci. 1994, 34, 1000-1008, Rubicon converts 87% of the structures, outperforming CONCORD v3.0.1 (84%), putting it behind CORINA in the benchmark and beyond most other programs (<79%).

  • Rubicon performance on internal benchmarks has improved. The conversion rate of the Interbioscreen99sc database is nearly 100%. The unprocessed cases are caused by organic elements, i.e., B and Si. Also, due to SMARTS pattern optimization, Rubicon is 18% faster on in-house tests.
  • Depiction Changes

    Some work has been done to make depictions more aesthetically pleasing. Specifically, long chains and large rings are layed out better, and single-atom molecules are displayed better.

  • A new depiction-object boolean property ("suppress") is provided. This property controls an intermediate display mode between schematic and non-schematic. In suppressed mode, long-chain carbons are not shown. Unlike schematic mode, aromatic rings and terminal atoms are shown. Xvthor and Xvmerlin now support the new option "DEPICT_SUPPRESS_LABELS" which controls this display mode in the applications. The default is to use the "suppress" mode.

  • System Administration Issues

    Upgrade path

  • Thor database file formats in 4.71 are identical to those in version 4.6x. There are several classes of SMILES for which the canonical representations change. The program conversion utility "thordbcheck471" is provided to identify and clean up these cases. See the man page for thordbcheck471 for usage instructions.

  • The separate Remote ToolkitTM is not shipped with version 4.71. The Remote Toolkit version 4.61 is 100% compatible with 4.61, 4.62 and 4.71 DayTool servers.
  • Supported Platforms

  • SUN Solaris 2.5, 2.6, & 2.7
  • SGI Irix 6.4 & 6.5
  • Red Hat Linux 6.1 & 6.2 (x86)

  • Daylight Chemical Information Systems, Inc.
    info@daylight.com