CAMERA is a new method for rapidly reconstructing NUS NMR data according to the statistical principle of maximum entropy. Using accelerated techniques for optimizing smooth convex functions, CAMERA converges to MaxEnt solutions faster than prior methods.
The acquisition of sparsely sampled time-domain data in NMR spectroscopy is becoming more and more popular, but most techniques for determining which Nyquist grid points to sample are based on pseudorandom numbers. The choice of which sample to select from the ensemble of possibilities is nontrivial, and is usually too tiresome for the practicing spectroscopist. This is my two cents on how to address the issue of seed-dependent sampling.
Pipelines for handling 1D NMR data for metabolomics and other chemometrics applications are typically either ad hoc toolchains of non-free software spiced with manual curation in Excel, or custom MATLAB toolkits that have never reached the public. My MVAPACK package for GNU Octave is meant to address the need for a FOSS 1D NMR chemometrics toolkit.
NMR datasets suffer from a phase-correction problem, where either a trained user or software algorithms must 'phase' spectra in order to extract highest quality information. Phase errors between spectra in a dataset have never been addressed, until I wrote pscorr, of course.
The bulk of the metabolomics community relies on visual inspection of PCA and PLS-DA scores plots to decide which experimental groups show significant differences. Because this is actually a simple statistical task and subject to human bias, I wrote a small set of C programs to rapidly assess the statistical significance of separations in multivariate scores.