Dear Bioconductor developers, I am happy to present my new package, available with the new version of Bioconductor. MEAL is a package designed to integrate methylation and expression and to analyze both separately. Integration of these data types can be univariate or multivariate. The univariate approach consists on two steps. First, cpgs and expression probes are paired depending on their genomic position. Afterwards, a linear regression is performed, considering the expression the outcome and the methylation the independent variable. On the other hand, the multivariate approach is designed to study the correlation between methylation and expression in a given genomic region. MEAL also can analyze methylation and expression separately. To do so, it includes the most common algorithms to analyze methylation and expression, including several region analysis such as Bumphunter or DMRcate. MEAL can also analyse methylation or expression in a given region, which could be of interest when studying genomic structural variants. Finally, it should be noticed that it includes the most common plots used in methylation and expression analysing, easing the process of presenting the results. The last feature to be mentioned, it is the development of a new class (MultiDataSet) designed to handle different omics sets of common samples. At the moment, it implements some methods to add MethylationSet (class of MEAL), ExpressionSet and SnpSet as well as a generic function to add any eSet derived object. This design will allow the reuse of the class to other integration tasks by other developpers. I hope this package will be very useful for people working with methylation and expression and that it could be a good starting point for other integration techniques. Bests, Carlos Ruiz, Center for Research in Environmental Epidemiology (CREAL)
[Bioc-devel] New package: MEAL - Methylation and Expression AnaLyzer
2 messages · Ruiz Arenas, Carlos, Nathan Sheffield
5 days later
We're pleased to announce the release of a new Bioconductor package, LOLA (Locus Overlap Analysis). LOLA is a package for genomic locus overlap enrichment. Roughly analogous to what GSEA does for gene sets, LOLA does for genomic regions, which can be defined however you like, including experiments like ChIP-seq, BS-seq, DNase-seq, etc. LOLA lets you test your genomic ranges of interest against a database of other genomic range sets to identify enrichment of overlap, tying external annotation to your regions of interest. A complete enrichment analysis against a database of thousands of region sets requires just 3 lines of R code and completes in minutes. Along with LOLA, we provide the LOLA Core Database, which includes region sets from ENCODE, Roadmap Epigenomics, Cistrome, CODEX, UCSC, and other public databases. We intend to build, maintain, and curate this database over the coming years. You can also use LOLA for custom analysis by creating custom databases following instructions in the LOLA readme. You can find more information, vignettes, and downloads at LOLA's website (http://databio.org/lola/), or follow our development at GitHub (https://github.com/sheffien/LOLA). We would be eager to hear any suggestions or feedback! Nathan Sheffield and Christoph Bock