Skip to main content
Fig. 3 | BMC Bioinformatics

Fig. 3

From: Constrained Fourier estimation of short-term time-series gene expression data reduces noise and improves clustering and gene regulatory network predictions

Fig. 3

Analysis of k-means clustering of raw (gray) and de-noised (red) synthetic expression data. A, B Total size (of all six clusters) of correlation and SSE between the raw signals to the true signals (gray) and de-noised signals to the true signals (red). C–F analyzes the performance of the clustering as a function of the number of data samples: C, D Total mean correlation and error (SSE) of expression signals from clusters of raw (gray) and de-noised (red) data, as a function of sampling frequency (linearly distributed). The difference is not statistically significant for over 7 sample points (two-sample t-test). E, F Mean correlation and SSE of clustering of raw (gray) and de-noised (red) data, as function of sampling frequency with a logarithmic time scale (see “Methods” section). The improvement in the clustering performance was significantly better over 5–7 sample points. And most importantly, above 7–8 samples the improvement is not statistically significant

Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us