Motivation: Methylation of cytosines in DNA plays an important role in the regulation of gene expression, and the analysis of methylation patterns is fundamental for the understanding of cell differentiation, aging processes, diseases and cancer development. Such analysis has been limited, because technologies for detailed and efficient high-throughput studies have not been available. We have developed a novel quantitative methylation analysis algorithm and workflow based on direct DNA sequencing of PCR products from bisulfite-treated DNA with high-throughput sequencing machines. This technology is a prerequisite for success of the Human Epigenome Project, the first large genome-wide sequencing study for DNA methylation in many different tissues. Methylation in tissue samples which are compositions of different cells is a quantitative information represented by cytosine/thymine proportions after bisulfite conversion of unmethylated cytosines to uracil and PCR. Calculation of quantitative methylation information from base proportions represented by different dye signals in four-dye sequencing trace files needs a specific algorithm handling imbalanced and overscaled signals, incomplete conversion, quality problems and basecaller artifacts. Results: The algorithm we developed has several key properties: it analyzes trace files from PCR products of bisulfite-treated DNA sequenced directly on ABI machines; it yields quantitative methylation measurements for individual cytosine positions after alignment with genomic reference sequences, signal normalization and estimation of effectiveness of bisulfite treatment; it works in a fully automated pipeline including data quality monitoring; it is efficient and avoids the usual cost of multiple sequencing runs on subclones to estimate DNA methylation. The power of our new algorithm is demonstrated with data from two test systems based on mixtures with known base compositions and defined methylation. In addition, the applicability is proven by identifying CpGs that are differentially methylated in real tissue samples.

We applied serial analysis of gene expression (SAGE) to study differentially expressed genes in mouse brain 14 hr after the induction of focal cerebral ischemia. Analysis of >60,000 transcripts revealed 83 upregulated and 94 downregulated transcripts (more than or equal to eightfold). Reproducibility was demonstrated by performing SAGE in duplicate on the same starting material. Metallothionein-II (MT-II) was the most significantly upregulated transcript in the ischemic hemisphere. MT-I and MT-II are assumed to be induced by metals, glucocorticoids, and inflammatory signals in a coordinated manner, yet their function remains elusive. Upregulation of both MT-I and MT-II was confirmed by Northern blotting. MT-I and MT-II mRNA expression increased as early as 2 hr after 2 hr of transient ischemia, with a maximum after 16 hr. Western blotting and immunohistochemistry revealed MT-I/-II upregulation in the ischemic hemisphere, whereas double labeling demonstrated the colocalization of MT with markers for astrocytes as well as for monocytes/macrophages. MT-I- and MT-II-deficient mice developed approximately threefold larger infarcts than wild-type mice and a significantly worse neurological outcome.For the first time we make available a comprehensive data set on brain ischemic gene expression and underscore the important protective role of metallothioneins in ischemic damage of the brain. Our results demonstrate the usefulness of SAGE to screen functionally relevant genes and the power of knock-out models in linking function to expression data generated by high throughput techniques.

A four-step procedure for the efficient and systematic mining of whole EST libraries for differentially expressed genes is presented. After eliminating redundant entries from the EST library under investigation (step 1), contigs of maximal length are built upon each remaining EST using about 4 000 000 public and proprietary ESTs (step 2). These putative genes are compared against a database comprising ESTs from 16 different tissues (both normal and tumour affected) to determine whether or not they are differentially expressed (step 3; electronic northern). Fisher's exact test is used to assess the significance of differential expression. In step 4, an attempt is made to characterise the contigs obtained in the assembly through database comparison. A case study of the CGAP library NCI_CGAP_Br1.1, a library made from three (well, moderately, and poorly differentiated) invasive ductal breast tumours (2126 ESTs in total) was carried out. Of the maximal contigs, 139 were found to be significantly (α = 0.05) over-expressed in breast tumour tissue, while 13 appeared to be down-regulated.