CAMPARI

March 7 2011 Version 1.0.1 is is a minor release specifically to fix a bug with the initialization of the pseudo-random number generator (PRNG) in parallel (MPI) runs. The intended behavior is to initialize the PRNG with different seeds for each replica. This did not necessarily happen in version 1.0 (depending on exact deployment time of each sub-job), but is fixed in 1.0.1 (bug was a result of inadvertent commenting of a few lines of code). The bug is unlikely to affect replica-exchange runs (http://campari.sourceforge.net/keywords.html#REMC), unless conditions where chosen to be quasi-identical. It could have introduced severe biases to MPI-averaging runs (http://campari.sourceforge.net/keywords.html#MPIAVG), however, due to the artificial overweighting of specific trajectories. Aside from cosmetic changes to some printed error messages, there is only one other notable change: a feature extension that now allows hydrogen bonds for the DSSP algorithm to also be considered if they involve polypeptide capping groups (previously, carbonly groups on N-terminal capping groups, and amide groups on C-terminal capping groups were rigorously ignored). July 8 2011 A very simple bug was found that would specifically break the computation of intermolecular distance histograms in parsing mode 1 of the general pair correlation analysis functionality (http://campari.sourceforge.net/inputfiles.html#FMCSC_PCCODEFILE). This bug had no effect if any of the type-generalized parsing modes was used (modes 2 or 3) or for intramolecular distance distributions even if mode 1 was used. Since the fix involves a single line of code, the corresponding source file (structure.f90) has been uploaded instead of a complete distribution. Users can also implement the fix themselves: Line 1472 of structure.f90 should be: pcbin = int(sqrt(dis2)/pc_binsz) + 1 Instead of: pcbin = int(dis/pc_binsz) + 1 With apologies, The CAMPARI Development Team November 4 2011 Two bugs in the documentation (specifically at http://campari.sourceforge.net/inputfiles.html#UAMODEL and http://campari.sourceforge.net/inputfiles.html#SCRMODEL) were fixed. In general, the online documentation will be safer to use for released features, since no new archive and no new files will be uploaded just to fix a couple of lines of documentation. Also please do not hesitate to contact us preferably via the SF-pages if there are features, improvements, etc. of interest. Current development versions (not posted here) may already incorporate them. With apologies, The CAMPARI Development Team February 24 2012 Another simple analysis bug was identified. This bug would specifically inflate the values reported for the mean-square radius of gyration in output file POLYAVG.dat (http://campari.sourceforge.net/outputfiles.html#POLYAVG.dat), if the underlying analysis did encompass more than one identical molecule. The factor by which the numbers are too large corresponds roughly to the number of identical molecules for the given analysis group, i.e., the bug should be fairly obvious to diagnose given the reported value for the mean radius of gyration in column 1. The error would also lead to incorrect values for asphericity and acylindricity (because of their normalization) reported in columns 4 and 5, but not for any other values in POLYAVG.dat. The proposed fix can be implemented by hand, and is documented in the corresponding Trac ticket (https://sourceforge.net/apps/trac/campari/ticket/3). Since the fix involves only a few lines of code, the affected source files (mod_polyavg.f90, polymer.f90, allocate.f90) have been uploaded as a zip-archive instead of a complete distribution. With apologies, The CAMPARI Development Team March 12 2012 A further simple analysis bug was identified that would incorrectly compute the average of chain dihedral angles reported in columns 3, 5, and 7 of output file TURNS_RES.dat (http://campari.sourceforge.net/outputfiles.html#TURNS_RES.dat). The problem was an incorrect unit when shifting computed averages (via the arctangent function) by constants of -180/180 degrees, respectively. Since these corrections were incorrectly applied in units of Radian, the resultant values were off by approximately +/-177 degrees. The bug had no effect on any other entry in TURNS_RES.dat. Those averages not requiring a shift were reported correctly as well. The proposed fix can easily be implemented by hand, and is documented in the corresponding Trac ticket (https://sourceforge.net/apps/trac/campari/ticket/4). The affected source file (polymer.f90) was previously uploaded as part of a zip-archive, and said archive has now been replaced with a version containing this newer fix as well. With apologies, The CAMPARI Development Team June 1 2012 A possible runtime bug was identified and corrected that could prevent successful restarts of dynamics simulations when using polypeptides with non standard (non-planar) geometries at the carbonyl carbon and amide hydrogen sites along the backbone. The error would manifest itself after the first step of dynamics after a restart, when the rebuilding of polypeptides would not account for these geometric peculiarities leading to a structural mismatch. The fix for this bug involves the addition of a single character to source file restart.f90, and can easily be implemented by hand (https://sourceforge.net/apps/trac/campari/ticket/5). Otherwise, a corrected version of the affected source file was uploaded to the SF servers, and is available for download. The CAMPARI Development Team January 15 2013 Another simple analysis bug was identified that would not use correct values for the entries of the gyration tensor when running a dynamics calculation (FMCSC_DYNAMICS being something different from 1). This bug affected results in output files RGHIST.dat, RDHIST.dat, and POLYAVG.dat, specifically for all quantities dependent on the entries of the gyration tensor, and was caused by a lack of updating those purely geometrical (not mass-dependent) quantities. The fix involves changes to one file (polymer.f90) that has played a role in two other analysis bugs already. It was previously uploaded as part of a zip-archive (02/12), said zip-archive was updated (03/12), and it is updated again now (meaning that the version of this zip-archive currently on SF provides fixes to all three bugs). This is also a good time to announce that the next major version of CAMPARI is expected to be released within the next few months (February to April). With apologies and best regards, The CAMPARI Development Team