Reversible jump MCMC approach for peak identification for stroke SELDI mass spectrometry using mixture model.

Mass spectrometry (MS) has shown great potential in detecting disease-related biomarkers for early diagnosis of stroke. To discover potential biomarkers from large volume of noisy MS data, peak detection must be performed first. This article proposes a novel automatic peak detection method for the stroke MS data. In this method, a mixture model is proposed to model the spectrum. Bayesian approach is used to estimate parameters of the mixture model, and Markov chain Monte Carlo method is employed to perform Bayesian inference. By introducing a reversible jump method, we can automatically estimate the number of peaks in the model. Instead of separating peak detection into substeps, the proposed peak detection method can do baseline correction, denoising and peak identification simultaneously. Therefore, it minimizes the risk of introducing irrecoverable bias and errors from each substep. In addition, this peak detection method does not require a manually selected denoising threshold. Experimental results on both simulated dataset and stroke MS dataset show that the proposed peak detection method not only has the ability to detect small signal-to-noise ratio peaks, but also greatly reduces false detection rate while maintaining the same sensitivity.

Knowledge-based computational intelligence development for predicting protein secondary structures from sequences.

In the postgenomic age, with the avalanche of protein sequences generated and relatively slow progress in determining their structures by experiments, it is important to develop automated methods to predict the structure of a protein from its sequence. The membrane proteins are a special group in the protein family that accounts for approximately 30% of all proteins; however, solved membrane protein structures only represent less than 1% of known protein structures to date. Although a great success has been achieved for developing computational intelligence techniques to predict secondary structures in both globular and membrane proteins, there is still much challenging work in this regard. In this review article, we firstly summarize the recent progress of automation methodology development in predicting protein secondary structures, especially in membrane proteins; we will then give some future directions in this research field.

[Automatic segmentation of three dimension medical image series].

We propose an improved version of regional competition algorithm in this paper, and apply it to the automatic segmentation of medical image series, particularly in the segmentation and recognition of brain tumor. The traditional regional competition is enhanced by combining the attractive aspects of fuzzy segmentation, and thus it provides an efficient approach to segment the fuzzy and heterogeneous medical images. In order to perform regional competition on medical image series, we utilize the segmentation result of a slice to initiate the next segmented slice, while the first slice is initialized using regional growing algorithm. Moreover, we develop an algorithm to recognize the tumors automatically, taking into account its characters. Experimental results show that our algorithm performs well on the segmentation of magnetic resonance imaging (MRI) image series with high speed and precision.