Sunday, July 29, 2007

SwaN-MR

Unbelievable. I can't understand. The last version of SwaN-MR dates March 2001. It does not run on today's computers. The web site was de facto closed in September 2005 (only a single page remained). In 2007 a snapshot of the old site (dated April 2000) reappeared as a sort of museum. Today, if you google "nmr mac" which site takes the highest rank? Try by yourself...
What puzzles me is another thing. Since November I am monitoring the clicks received by the only visible link on the only remaining page. There are clicks every day. People with nothing better to do or people in need of software? It is possible to collect some more information, like country, IP, browser, etc..
Most of the times the remnant of the SwaN-MR web site is visited by somebody using Windows. Does it mean that thousand of old SwaN-MR users have migrated to Windows, but are still affected by nostalgia? Does it mean anything else? Now, PLEASE, if you recognize yourself as the subject of this post, would you mind to leave a short comment?

Friday, July 20, 2007

iNMR

I wrote this review 6 weeks ago for MacResearch.org. A couple of sentences have rapidly become outdated. I have rephrased them and now they are ready to stand the test of time. For how long?

Nuclear Magnetic Resonance has been introduced in chemistry about 50 years ago and its importance has exponentially increased up to the point that an NMR spectrum is required to characterize any new compound. This spectroscopic technique is also widely used in biochemistry, medicine and other fields, but chemists still represent the majority of the users. Since the sixties a variety of software has been written to work with NMR data; initially to extract chemical shifts and coupling constants from overcomplicated CW spectra, then to perform the Fourier Transform on pulsed NMR spectra, successively to process 2D spectra of bio-polymers; in recent years there has been a return of interest towards the analysis of medium size molecules.

The population of chemists can be divided into three categories: 1) those who can't understand the need of processing NMR spectra on a personal computer and stick to the hardware and software every spectrometer comes equipped with; 2) those who have limited access to the spectrometer or need to process the spectra in the quiet of their rooms; 3) those who are not satisfied by the software furnished with the instrument and look for alternatives.

Until the 90s it was impossible to process an FT-NMR spectrum on personal computers, both because it was difficult to transfer and because only workstations were powerful enough. By the end of the century the situation had changed. On the Macintosh the main program for the task was SwaN-MR and, being it freely available on the internet and frequently updated, it became universally popular.

With the arrival of OS X there have been a few years in which the only option was to run SwaN-MR in classic mode. In 2005 things began to change. First Varian ported VNMRJ from Solaris, but this program has never become popular on any platform, and today its use is confined to the spectrometers consoles. In the next summer the Java-based "NMR NoteBook" appeared but it stopped forever at version 1.0. Today we have Jeol Delta, which has nothing in common with the Aqua interface and is difficult to learn, the unix-based NMRpipe, which has been in use for a long time but whose scope is restricted (namely: multidimensional spectra for conformational analysis of proteins) and iNMR, the subject of this review. In June Mestrelab Research has brought in a concurrent program, called MestreNova, which is a cross-platform product. It is based on the QT library which, with all its strengths, presents some interface limitations on the Mac platform. For example you are forced to work into a "child window" (a familiar concept in the Windows environment but out of place on the mac) and some commands (e.g.: the application preferences) are not where they seems to be.

The penury has come to an end, the options are many, but only one of them can be considered an authentic Mac application from all points of view. It naturally becomes the first option to consider when a Mac user looks for an NMR program.

iNMR has other peculiarities, besides being an OS X product. The author combines 20 years of experience both in programming and in using NMR software for research activities. The consequences are evident when you start working with iNMR and realize how every detail has been carefully thought, and how many ancient issues have been solved with innovative solutions. A third peculiarity is that iNMR does not create yet another file format. All other NMR applications require, as a preliminary step, the translation of external files into their own format. iNMR works, instead, directly with the original data sets. These feature allows the existence of a correlated program, "iNMR reader", which can be approximatively described as iNMR deprived of the "Save" command. The reader is popular among students because of its affordable price. The most interesting fact, however, is that iNMR users don't have to worry about future compatibility issues, because the original files, those coming from the spectrometer, remain untouched and they are already recognized by other software on all platforms.

The public targeted by iNMR is composed by curious researchers. Even if almost all commands are scriptable and can be automated (all the following tasks are automatically performed: recognition of the format of the incoming files, Fourier Transform, Phase and Baseline Correction, TMS recognition, Integration) there is however a strong stress on interactivity. Some ideas seem to be taken from the world of games. For example: the screen is completely dedicated to the plot of the spectrum (you see no toolbar, no rulers, no scroll-bars, no tabs, etc... only the ppm scale and a palette, both optional). The user feels to be fully immersed into the data. Second example: the extreme interactivity. On many an occasion a parameter is not entered into a text field, but controlled through a slider. While the slider moves, the plot is updated in real time. The effect is pleasantly smooth, because iNMR is apparently well optimized for speed; it has a small memory imprint too (only 3 MB, and it's Universal Binary!). Interactivity is also the main ingredient of the simulation and deconvolution modules, in which the user can grab single peaks with the mouse and move or resize them.
A researcher that needs to study her/his spectra for a prolonged time will certainly welcome this kind of "playability". You can really avoid printing, because all the details are perfectly visible on the screen.

This review comes without pictures, but the official site (http://www.inmr.net) offers numerous screenshots and the program itself can be freely downloaded and tried without time limitations (but some commands are disabled before registration). There is an associated open discussion group (http://groups.google.com/group/inmr) generically dedicated to "NMR & Mac". Members receive a free copy of iNMR 0.7, so in practice every one can receive a free copy.

The processing section includes all the commonly used algorithms plus such advanced options as Difference Spectroscopy, Reference Deconvolution, Direct Editing of the data points, Solvent Suppression, Noise Removal, Linear Prediction, Hilbert Transform, Binning and Tilting.
There are many interesting modules embedded into iNMR which will be only listed here: line-fitting of overlapping peaks (usually called "deconvolution" by NMR spectroscopists), simulations based on quantum-mechanics (including general cases of chemical exchange), a database empowered by Spotlight (you can use it as a smarter substitute of the Open dialog), a console for the Lua programming language (call it a command line on steroids), tools for annotating the plot (with text, arrows, brackets, frames and clip-art) and a growing number of facilities for metabolomic analysis. Among these a proprietary algorithm for automatic phase correction that is claimed to be "the most accurate on the market". It is also possible to export the processed data in a variety of formats, like pictures, text, binary files or JCAMP-DX (the standard for exchanging spectra). One impressive example is the command "Copy Integrals" that creates a comparative table with the integrals of a series of spectra. The advantage for the user is that the integration intervals must be specified for the first spectrum only.

iNMR is a vast application and the number of features is continuously growing (there is at least an update each month). Probably nobody will ever find the time to explore all the corners. The NMR field is however unlimited and any application inevitably lacks some important feature. I have used or tested more than a dozen of similar applications and I have always seen the rule confirmed. It will always be, because not every user needs every feature.
The interested reader should check the program directly, or consult the feature list or contact the maker or the discussion group. iNMR is however highly usable, stable and dependable. Any laboratory can safely invest the modest sum that is required to buy a copy. If somebody wants an NMR program with an authentic Mac look & feel, no other option is available today and in the foreseeable future.