Adaptive kernel fuzzy clustering for missing data.

Many machine learning procedures, including clustering analysis are often affected by missing values. This work aims to propose and evaluate a Kernel Fuzzy C-means clustering algorithm considering the kernelization of the metric with local adaptive distances (VKFCM-K-LP) under three types of strategies to deal with missing data. The first strategy, called Whole Data Strategy (WDS), performs clustering only on the complete part of the dataset, i.e. it discards all instances with missing data. The second approach uses the Partial Distance Strategy (PDS), in which partial distances are computed among all available resources and then re-scaled by the reciprocal of the proportion of observed values. The third technique, called Optimal Completion Strategy (OCS), computes missing values iteratively as auxiliary variables in the optimization of a suitable objective function. The clustering results were evaluated according to different metrics. The best performance of the clustering algorithm was achieved under the PDS and OCS strategies. Under the OCS approach, new datasets were derive and the missing values were estimated dynamically in the optimization process. The results of clustering under the OCS strategy also presented a superior performance when compared to the resulting clusters obtained by applying the VKFCM-K-LP algorithm on a version where missing values are previously imputed by the mean or the median of the observed values.

Epigenetic age predictions based on buccal swabs are more precise in combination with cell type-specific DNA methylation signatures.

Aging is reflected by highly reproducible DNA methylation (DNAm) changes that open new perspectives for estimation of chronological age in legal medicine. DNA can be harvested non-invasively from cells at the inside of a person's cheek using buccal swabs - but these specimens resemble heterogeneous mixtures of buccal epithelial cells and leukocytes with different epigenetic makeup. In this study, we have trained an age predictor based on three age-associated CpG sites (associated with the genesPDE4C, ASPA, and ITGA2B) for swab samples to reach a mean absolute deviation (MAD) between predicted and chronological age of 4.3 years in a training set and of 7.03 years in a validation set. Subsequently, the composition of buccal epithelial cells versus leukocytes was estimated by two additional CpGs (associated with the genes CD6 and SERPINB5). Results of this "Buccal-Cell-Signature" correlated with cell counts in cytological stains (R2 = 0.94). Combination of cell type-specific and age-associated CpGs into one multivariate model enabled age predictions with MADs of 5.09 years and 5.12 years in two independent validation sets. Our results demonstrate that the cellular composition in buccal swab samples can be determined by DNAm at two cell type-specific CpGs to improve epigenetic age predictions.

DNA methylation levels at individual age-associated CpG sites can be indicative for life expectancy.

DNA-methylation (DNAm) levels at age-associated CpG sites can be combined into epigenetic aging signatures to estimate donor age. It has been demonstrated that the difference between such epigenetic age-predictions and chronological age is indicative for of all-cause mortality in later life. In this study, we tested alternative epigenetic signatures and followed the hypothesis that even individual age-associated CpG sites might be indicative for life-expectancy. Using a 99-CpG aging model, a five-year higher age-prediction was associated with 11% greater mortality risk in DNAm profiles of the Lothian Birth Cohort 1921 study. However, models based on three CpGs, or even individual CpGs, generally revealed very high offsets in age-predictions if applied to independent microarray datasets. On the other hand, we demonstrate that DNAm levels at several individual age-associated CpGs seem to be associated with life expectancy - e.g., at CpGs associated with the genesPDE4C and CLCN6. Our results support the notion that small aging signatures should rather be analysed by more quantitative methods, such as site-specific pyrosequencing, as the precision of age-predictions is rather low on independent microarray datasets. Nevertheless, the results hold the perspective that simple epigenetic biomarkers, based on few or individual age-associated CpGs, could assist the estimation of biological age.

Epigenetic Classification of Human Mesenchymal Stromal Cells.

Standardization of mesenchymal stromal cells (MSCs) is hampered by the lack of a precise definition for these cell preparations; for example, there are no molecular markers to discern MSCs and fibroblasts. In this study, we followed the hypothesis that specific DNA methylation (DNAm) patterns can assist classification of MSCs. We utilized 190 DNAm profiles to address the impact of tissue of origin, donor age, replicative senescence, and serum supplements on the epigenetic makeup. Based on this, we elaborated a simple epigenetic signature based on two CpG sites to classify MSCs and fibroblasts, referred to as the Epi-MSC-Score. Another two-CpG signature can distinguish between MSCs from bone marrow and adipose tissue, referred to as the Epi-Tissue-Score. These assays were validated by site-specific pyrosequencing analysis in 34 primary cell preparations. Furthermore, even individual subclones of MSCs were correctly classified by our epigenetic signatures. In summary, we propose an alternative concept to use DNAm patterns for molecular definition of cell preparations, and our epigenetic scores facilitate robust and cost-effective quality control of MSC cultures.