Skip to content

[Bioc-devel] Manifest and annotation packages for EPIC methylation array v2.0

3 messages · Shepherd, Lori, Zuguang Gu

Original
Zuguang Gu
# 
Hi Bioc-devel,

I recently analyzed a methylation array dataset which is from EPIC array
v2.0 (
https://www.illumina.com/products/by-type/microarray-kits/infinium-methylation-epic.html)
and currently there is no manifest nor annotation packages for that one.
The available packages on Bioconductor are only for EPIC array v1.0:

- IlluminaHumanMethylationEPICmanifest
- IlluminaHumanMethylationEPICanno.ilm10b4.hg19

And it is not proper to use the v1 packages for the v2 data.

I've built a manifest package and an annotation package for the EPIC array
v2.0. They are available at:

https://github.com/jokergoo/IlluminaHumanMethylationEPICv2manifest
https://github.com/jokergoo/IlluminaHumanMethylationEPICv2anno.20a1.hg38

Basically they are built with the scripts in the original v1 packages, with
small adjustment. The scripts for building the two packages are also
included in the two packages.

The two v2 packages can be seamlessly integrated into the minfi analysis.

I also did some simple comparisons between v1 and v2 (
https://jokergoo.github.io/IlluminaHumanMethylationEPICv2manifest/articles/IlluminaHumanMethylationEPICv2manifest.html
).

Since the v1 packages were developed by the bioc core members, I want to
ask is there any plan to support the EPIC v2 array officially? If there is
no such plan, can I submit the two packages to bioc? If I can, then I will
have some additional questions regarding how to pass BiocCheck (e.g. file
size, lazy loading, ...). But that will be in another email.

Best,
Zuguang
2 days later
Shepherd, Lori
# 
The maintainer for these two packages

IlluminaHumanMethylationEPICmanifest
IlluminaHumanMethylationEPICanno.ilm10b4.hg19

Is actually Kasper Daniel Hansen as listed in the DESCRIPTION file not bioc core members.

You could reach out to Kasper to see if there is any intention or efforts that have already been made to include v2.

If not you can submit the packages. We have moved away from traditional annotation packages and I would suggest moving towards an annotationhub package submission where the data is stored elsewhere on a external server like S3 buckets, Azure Data Lakes, Zenodo, etc.... (we have an bioconductor azure data lake location that can be used;  it just cannot be store on any personal location like github or dropbox)
Little modification to an existing package structure is needed. Basically you would remove the data to the external location,  include a metadata.csv file in the inst/extdata, and include the biocView AnnotationHub. More details can be found at
https://bioconductor.org/packages/release/bioc/vignettes/HubPub/inst/doc/CreateAHubPackage.html
Then submit as normal to the new package submission tracker.


Cheers,



Lori Shepherd - Kern

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263
Zuguang Gu
# 
I am aware of the AnnotationHub package, but currently I think it is not
possible to integrate with minfi.

In the last several lines in
https://github.com/hansenlab/minfi/blob/master/R/IlluminaMethylationAnnotation-class.R

setMethod(
    "getManifest",
    signature(object = "IlluminaMethylationAnnotation"),
    function(object) {
        maniString <- .getManifestString(object at annotation)
        if (!require(maniString, character.only = TRUE)) {
            stop(sprintf("cannot load manifest package %s", maniString))
        }
        get(maniString)
    }
)

Here `maniString` contains the name of a manifest package,
e.g. IlluminaHumanMethylationEPICmanifest. After loading the manifest
package, minfi directly looks for the variable with the same name as the
manifest package name. So it would be the easiest if just making a new
manifest package taking the IlluminaHumanMethylationEPICmanifest as a
template. Or minfi needs to be adjusted to allow for more flexible
integration of new manifest packages/objects.

Cheers,
Zuguang

On Tue, 21 Mar 2023 at 14:58, Kern, Lori <Lori.Shepherd at roswellpark.org>
wrote: