Version 1.4.2 has been released today. The release contains a few minor bugfixes and a minor revision to the scoring algorithm. Most people are not going to notice any difference in performance. The documentation has also been further updated to include more details on Percolator usage and interpretation. In the next update, I hope to add FAQ and troubleshooting sections for some of the most common issues observed.
I am going to make an effort to provide monthly updates. These might not necessarily be new releases, just development reports. In this case it is a new release. The latest release is version 1.4.1, which is mostly minor bug fixes. Most users are likely not affected. As always, for those who want bleeding edge, you can grab the latest source ahead of official releases from the source page.
A new version of Kojak was pushed up to the repository last friday. The compiled versions are now available for download. New features include addition of other ion series to the analysis and batch mode for multiple file analysis using a single set of parameters.
The alternate ion series allow for a, b, c, x, y, and z ion series to be included or excluded from the analysis as desired. With these features, ETD analyses can now be performed with cross-links. To facilite these analyses, the z-ion series are actually z-dot ions. I don’t have any cross-linked ETD spectra (only non-linked), so my test dataset is somewhat lacking. If anyone is willing to share some cross-linked ETD data, please send me a message.
Additionally, batch analysis was added to Kojak. Batch analysis allows for multiple files to be queued for analysis using a single set of parameters and a single command line execution. Use of batch analysis is simple: on the command line, after specifying your configuration file, list all files to be analyzed in the batch (wildcards accepted where supported). Kojak can still be used in single mode, specifying the file to be analyzed with the MS_data_file parameter. Files exported from Kojak will now be named from the input file name, plus a small text tag appended to the file. Thus, it is no longer necessary to provide an output file name, which is problematic when batching multiple files. Therefore, the output_file and percolator_file parameters have been deprecated. Exporting Percolator-ready results can be toggled on and off using the new export_percolator parameter. Please see the parameter list online for further details.
Thanks to everyone who was able to test and report the failure of the 1.3.6 Windows 64-bit Kojak installer. Sorry I was out of town and unable to fix it quickly. I uploaded a new version of the installer, switching to .msi format. I hope this resolves the issue. One possible problem is that I cannot test if the .msi is correctly installing the runtime environment, as I don’t have a test box without that environment installed. You will know if it is missing when you run Kojak and receive a missing driver error. Please send me a message if you have this problem.
A new version of Kojak was released yesterday, with many bug fixes, and some new features as well. See the release notes for full details. Some of the highlights include improved precursor prediction functionality. Part of the improvements give the option of suggesting multiple precursor masses when the selected mass window has poor quality signals. Though this adds some more computation time, it leaves fewer spectra unanalyzed.
Additionally, a few users were kind enough to report problems when analyzing, specifically, Thermo Fusion data on 64-bit Windows boxes. The cause appears to be some divergence in the Thermo driver functionality between 32-bit and 64-bit Windows systems. A workaround was put into place and it appears Fusion data can now be analyzed correctly on both 32-bit and 64-bit Windows systems. On a side note, if Kojak is unable to detect your 32-bit Thermo MSFileReader drivers, downgrading to driver version 2.2 solves this problem. A more appropriate solution may become available in the near future.
Thanks everyone who swung by my poster at ASMS and said hello! It was great to meet and talk to so many users, and I appreciate the comments you shared. The most common feature request is support for ETD, which I am working on implementing in the next version.
A few people have been kind enough to contact me after experiencing a missing dll error when using the 64-bit Windows version of Kojak. The error comes from missing the latest Microsoft VC redistributable package. Microsoft provides the package for free on their website, but typically it gets installed on a computer automatically with 3rd party software that requires it. Because I’ve been distributing Kojak in a zip file (i.e. not installed), the missing dll would not be installed for you. If you’ve been able to run 64-bit Windows Kojak without any errors, it was because some other software had already installed the drivers on your system.
The solution is to provide Kojak packaged into a setup/installation program. Please don’t panic when a window opens up asking if you would like to install the Microsoft Visual C++ Runtime Libraries. This is normal behavior and you must agree for Kojak to work on your computer. I’ve provided this installer for the 64-bit Windows Kojak version only (so far), as the 32-bit version uses older, well established drivers, and the linux version doesn’t need them at all. The caveat to using a setup program is that it requires administrative privileges - if you don’t have them and try installing Kojak, you will receive an error message. Removing Kojak after using the installer should be done from the control panel. The Kojak zip file will still be provided for people who prefer that method.
Another driver related issue that has appeared on 64-bit Windows systems is getting Thermo RAW files to read natively in Kojak. If you have the 32-bit MSFileReader driver from Thermo installed on your 64-bit Windows box, please uninstall it and install the 64-bit version. This corrects the problem.
In addition to continual document updates, I will soon be adding a Troubleshooting and FAQ section, compiled from user communications. Thank you very much to everyone who has written to let me know about any issues they encountered (and the success stories). Your feedback is greatly appreciated!
A new release of Kojak was pushed out today. Thanks everyone for the feedback and support, it is helping to make the algorithm even stronger.
The new version includes a new cross-linker reactive group, amine+. This is the same as the amine reactive group, but adds the rare alternative sites at serine, threonine, and tyrosine. If used with an amine-amine homobifunctional cross-linker, I recommend combining only a single amine+ to amine, as it is exceptionally unlikely to see, for example, serine-serine cross-links this way, and not worth the additional computation time.
Also the intra/inter file naming has been corrected for real this time. Sorry that it did not make it into the last version as originally stated. I think the code fix got lost in the migration from Google Code to GitHub.
The site is still being developed with unannounced updates about once a week. A few new pages were recently added. Next up is a tutorial on using Percolator. Stay tuned.
The really good news is that the Journal of Proteome Research accepted the Kojak manuscript yesterday. The not as good, but still nice, news is that the new website template is live.
Regarding the publication, I’ve added a Citing & Publications section to the Documentation. It contains the appropriate reference for citing Kojak, and list references of research projects that have used Kojak. Are you using Kojak? Contact me and I will reference your publication.
As for the new site layout, it looks superficially like I’ve merely added a template to make it marginally prettier. Actually, under the hood it has a lot more code so that I can continue to develop the resources provided on this site. I expect to have the Instructions fleshed out much quicker now. Stay tuned.
As a final note, the source repository for Kojak has been updated to include the missing source file. That file, as explained on the Source page, was omitted to protect the integrity of the review process and prevent forks of Kojak appearing everywhere while the reviewers were still conducting their scrutiny. Now that the source file has been included, Kojak should compile without errors. Feel free to fork and develop the source to tailor it to your needs. Similarly, the password protection on the compiled downloads is being removed, as these are no longer restricted to the reviewers for use.
Here’s the new and improved website! A simple template was applied to add a splash of color and the front page was redesigned with more flair. The rest of the changes are all under the hood, simply to make my life easier.