Implement support for methylation data
|Reported by:||Nicklas Nordborg||Owned by:||Nicklas Nordborg|
Description (last modified by )
This ticket is about importing (EPIC) methylation data from tumor DNA samples. It is somewhat similar to the OncoArray SNP (see #1346) data in the structure.
The starting point is DNA items in BASE. We need to make sure that there are aliquots created that represents the amount taken from the DNA item. Typical name will be something like
1234567.1.l.d.aX (where X may be missing or a number). The samples sheets that we have only have the
1234567 as identifier so we need to map this first.
The DNA aliquot should be linked to
PhysicalBioAssay item representing the actual EPIC chip that was used. It
might is be possible to re-use the
BeadChip suptype for this with but with new value for the
BeadChipType annotation (eg.
The other alternative is to use a new subtype (eg. This item also need the
BeadChipID allows us to find the data files related to the chip, since they are all contained in a single sub-directory. Each chip has 8 locations which should be linked to the 8 DNA aliquots used on the chip. The DNA aliquots need to have this information as annotations (
Scanning of the chip is represented by a DerivedBioAssay
Scan item. Dates, Scanner ID, etc. can be extracted from the data we have and should be imported as annotations or linked Hardware items. The
DataFilesFolder annotation will point to a folder with the scan data. We will need the *.idat files in the next step.
The scan will produce one pair of *.idat files for each of the 8 samples on the chip. We will represent this as 8 child DerivedBioAssay items with a new subtype
Methylation. The *.idat files are are stored in the
<project-archive> using the existing naming convention (eg.
We want to do some analysis that produces beta-values. This will be part of another ticket.
See #1419 for a wizard that calculate beta-values.