Wavelet analysis of electron-density maps

The wavelet transform is a powerful technique in signal processing and image analysis and it is shown here that wavelet analysis of low-resolution electron-density maps has the potential to increase their resolution. Like Fourier analysis, wavelet analysis expresses the image (electron density) in terms of a set of orthogonal functions. In the case of the Fourier transform, these functions are sines and cosines and each one contributes to the whole of the image. In contrast, the wavelet functions (simply called wavelets) can be quite localized and may only contribute to a small part of the image. This gives control over the amount of detail added to the map as the resolution increases. The mathematical details are outlined and an algorithm which achieves a resolution increase from 10 to 7 Å using a knowledge of the wavelet-coefficient histograms, electron-density histogram and the observed structure amplitudes is described. These histograms are calculated from the electron density of known structures, but it seems likely that the histograms can be predicted, just as electron-density histograms are at high resolution. The results show that the wavelet coefficients contain the information necessary to increase the resolution of electron-density maps.