Comet release 2016.01
release 2016.01 rev. 2 (2016.01.2), release date 2016/04/06
- Reverts the modification encoding in the "output_txtfile" output back
to reporting variable modification mass differences (e.g. DLSTM[16.0]HK)
instead of the actual modified residue mass values (e.g. DLSTMHK).
- Known bug: there's an issue with raw file parsing where the charge state
and precursor mass might not be correctly read. Mike Hoopmann made an
update to the MSToolkit file parsing library to fix this. I've decided
not to make a new release (branching and tagging the code and assigning a
new version string). Please download the comet_source_2016012.zip or
comet_binaries_2016012.zip files again to obtain the patched program.
These zip files were updated on 10/13/2016. Thanks to A. Sharma for
reporting the bug.
release 2016.01 rev. 1 (2016.01.1), release date 2016/03/29
- Fixes a bug where variable terminal modifications on decoy peptides are not properly generated
leading to possible program crash due to calculating fragment ions that are too large for the
decoy peptide. Reported by Villen lab.
- Fixes a bug where peptide mass comparisons are failing, leading to duplicate peptides being
stored and reported as separate entries. Reported by S. Michalakopoulos.
- For text and pepXML output, the peptide rank for ties (two peptides with exact same xcorr)
now have the same rank instead of the rankings always being sequential.
- The 2016.01 release also changed how modifications are reported in the text output. Instead
of mass differences (e.g. DLSTM[16.0]HK), the modification annotation now shows the
modified residue mass (e.g. DLSTMHK).
release 2016.01 rev. 0 (2016.01.0), release date 2016/03/15
- Known bug: when using a decoy search and variable n- or c-terminal modifications,
Comet could segfault/crash due to accessing an array out of bounds. This is due to
the modfications not being properly translated to the decoy peptide which could result
in calculated fragment ion masses that are too large. A fix is implemented and will
be part of the next release. Until the next release is available, email me if you
would like a patched binary.
- Add direct selenocysteine support by retiring parameter entry "add_U_user_amino_acid" and adding
- Allow negative numbers for the
parameter; will subtract that many threads from # CPU cores. For example, to use 3 threads for a 4-core
CPU or 7 threads for an 8-core CPU, set "num_threads = -1". This allows your computer to have a
CPU core free when running searches. Requested by D. Shteynberg.
- Fix bug in the expectation score calculation when
"print_expect_score = 0"
and pep.xml, pin, and txt files are specified (but not .out files). Under this scenario, the cross
correlation histogram was not being fully accumulated which is what is used to calculate the E-value.
This fix should result in more accurate E-values. Reported by A. Cheng.
was intended to only affect ".out" output.)
- For Comet compiled for Crux,
return the missing modified peptide column in the text output.
- Update deltaCn calculation for non top-ranked peptides or .pep.xml, .pin and .txt outputs.
The deltaCn values for the top ranked hits do not change. The lower hit entries had incorrect
values associated with them (normalized xcorr difference between consecutive entries instead
of the normalized xcorr difference from the top hit, the latter which is correct). Reported by M. Hoopmann.
- This version of Comet will run with comet.params files generated from version 2015.02. If you
use these parameter files, you will receive a warning message about an unknown parameter
"add_U_user_amino_acid" that you can safely ignore. Set "add_U_user_amino_acid = 150.95363" for
selenocysteine support with the old parameters file.
Documentation for parameters for release 2016.01
can be found here.
Go download from the download page.