'Annotation and Visualisation of Differential Expression Results' explains there are multiple ways to obtain annotations and then demonstrates how to do this with AnnotationHub, which seems perfectly reasonable.
|
Today we will use the `AnnotationHub` method. A workflow for annotation with |
|
biomaRt is included in the extended materials section accessible on the course |
|
website. |
However, after this, we then load a set of annotations that was obtained via biomaRt. I assume this is a legacy issue. This is a bit confusing and I would suggest the loaded annotations should be obtained with AnnotationHub too.
|
## One we prepared earlier and one-to-many relationships |
|
|
|
|
|
To ensure everyone is working with same annotation, we have created an annotation table. |
|
|
|
In this case we used the `biomaRt` package to download annotations directly from |
|
Ensembl. In this cases we can get additional columns, but will also sometimes get |
|
one-to-many relationships, where one Ensembl ID maps to multiple Entrez IDs. This |
|
sort of problem is common when mapping between annotation sources, but they have |
|
already been dealt with for us in AnnotationHub. If we wanted more control over |
|
this we would need to manually curate these one-to-many relationships ourselves. |
|
|
|
In annotation table below we have modified the column names and dealt with the |
|
one-to-many/missing issues for Entrez IDs. The code we used for doing this is |
|
available in the [extended materials section](S6_Annotation_With_BioMart.html). |
|
|
|
We will load out pre-created annotation table, and then combine it with our |
|
results table. |
|
|
|
```{r addAnnotation, message=FALSE} |
|
ensemblAnnot <- readRDS("RObjects/Ensembl_annotations.rds") |
