Question

fail to annotate chromosome data to single cell object

0

Entering edit mode

sandra.garces.19 • 0

@fc71d99c

Last seen 13 months ago

United Kingdom

I am trying to annotate my genes with ensemble and

ah = AnnotationHub()
ens.hs.107<- query(ah, c("Homo sapiens", "EnsDb", 107))[[1]] 

genes <- rowData(sce)$ID
head(genes)

gene_annot <- AnnotationDbi::select(ens.hs.107, 
                                    keys = genes,
                                    keytype = "GENEID",
                                    columns = c("GENEID", "SEQNAME")) %>%
  set_names(c("ID", "Chromosome"))

head(gene_annot)

rowData(sce) <- merge(rowData(sce), gene_annot, by.= "ID", sort=FALSE)
Error in .local(x, ..., value = value) : 
  26554 rows in value to replace 26664rows
##  genes found in genne_annot  have fewer GENEIDs than genes in sce object
rownames(rowData(sce)) <- rowData(sce)$ID

I have tried to use by.x and all.x but not sure its correct as I don't know why I have fewer geneids

 by.x, all.x= T
#and
 by.x,

Thank you!

error SingleCell annotate • 1.5k views

ADD COMMENT • link updated 12 months ago by yw24418 • 0 • written 20 months ago by sandra.garces.19 • 0

score 0 · Answer 1 · 2023-03-10

0

Entering edit mode

James W. MacDonald 67k

@james-w-macdonald-5106

Last seen 1 day ago

United States

I know nothing of the tidyverse, so can't help with how you are doing things with that magrittr pipe. But it's easy enough using base R functions. Also note that you don't have to ask for the GENEID to be returned as a column when using select, as that happens by default.

gene_annot <- select(ens.hs.107, rowData(sce)$ID, "SEQNAME", "GENEID")
rd <- rowData(sce)
rd$Chr <- gene_annot[match(rd$ID, gene_annot$GENEID),"SEQNAME"]
rowData(sce) <- rd

ADD COMMENT • link 20 months ago James W. MacDonald 67k

0

Entering edit mode

Thank you!

I have retrieved chromosomes but still have the same problem with the length of ensbl ID and gene ID

> dim(gene_annot)
[1] 26634     2
> dim(sce)
[1] 26664 65897

ADD REPLY • link 20 months ago sandra.garces.19 • 0

0

Entering edit mode

What do you get from

all(rowData(sce)$ID %in% keys(ens.hs.107, "GENEID"))

You might have mismatched versions.

ADD REPLY • link 20 months ago James W. MacDonald 67k

0

Entering edit mode

Or it might be PAR genes.

> gns <- genes(ens.hs.107)
Warning message:
In valid.GenomicRanges.seqinfo(x, suggest.trim = TRUE) :
  GRanges object contains 1 out-of-bound range located on sequence
  LRG_432. Note that ranges located on a sequence whose length is unknown
  (NA) or on a circular sequence are not considered out-of-bound (use
  seqlengths() and isCircular() to get the lengths and circularity flags
  of the underlying sequences). You can use trim() to trim these ranges.
  See ?`trim,GenomicRanges-method` for more information.

> ypar <- GRanges(c("Y","Y"), IRanges(c(1,56887903),c(10000,57217415)))
> subsetByOverlaps(gns, ypar)
GRanges object with 0 ranges and 9 metadata columns:
   seqnames    ranges strand |     gene_id   gene_name gene_biotype
      <Rle> <IRanges>  <Rle> | <character> <character>  <character>
   seq_coord_system description gene_id_version canonical_transcript
        <character> <character>     <character>          <character>
        symbol entrezid
   <character>   <list>
  -------
  seqinfo: 457 sequences (1 circular) from GRCh38 genome

I know that Ensembl hard masks the Y PAR region in their genome FASTA files, and maybe they do the same for the genes that occur in those regions? UCSC doesn't appear to.

> library(TxDb.Hsapiens.UCSC.hg38.refGene)
> gns2 <- genes(TxDb.Hsapiens.UCSC.hg38.refGene, single.strand.genes.only = FALSE)
> ypar2 <- ypar
> seqlevelsStyle(ypar2) <- "UCSC"
> subsetByOverlaps(gns2, ypar2)
GRangesList object of length 5:
$`100128260`
GRanges object with 2 ranges and 0 metadata columns:
      seqnames              ranges strand
         <Rle>           <IRanges>  <Rle>
  [1]     chrX 156014564-156016830      -
  [2]     chrY   57201084-57203350      -
  -------
  seqinfo: 640 sequences (1 circular) from hg38 genome

$`10251`
GRanges object with 2 ranges and 0 metadata columns:
      seqnames              ranges strand
         <Rle>           <IRanges>  <Rle>
  [1]     chrX 155612565-155782457      +
  [2]     chrY   56954255-56968977      +
  -------
  seqinfo: 640 sequences (1 circular) from hg38 genome

$`3581`
GRanges object with 2 ranges and 0 metadata columns:
      seqnames              ranges strand
         <Rle>           <IRanges>  <Rle>
  [1]     chrX 155997696-156010817      +
  [2]     chrY   57184216-57197337      +
  -------
  seqinfo: 640 sequences (1 circular) from hg38 genome

$`6845`
GRanges object with 2 ranges and 0 metadata columns:
      seqnames              ranges strand
         <Rle>           <IRanges>  <Rle>
  [1]     chrX 155881345-155943769      +
  [2]     chrY   57067865-57130289      +
  -------
  seqinfo: 640 sequences (1 circular) from hg38 genome

$`727856`
GRanges object with 2 ranges and 0 metadata columns:
      seqnames              ranges strand
         <Rle>           <IRanges>  <Rle>
  [1]     chrX 156025658-156028183      -
  [2]     chrY   57212178-57214703      -
  -------
  seqinfo: 640 sequences (1 circular) from hg38 genome

ADD REPLY • link 20 months ago James W. MacDonald 67k

0

Entering edit mode

It gives

False

I have found cell ranger used GRCh38-2020-A for mapping so I am using ensbl 101 now. I have 30 GENEIDs that don't match, maybe is the Y PAR genes but just wonder why more people don't have the same issue. Not sure how to get around it and/or whether this will have an effect on the downstream analysis.

ADD REPLY • link 20 months ago sandra.garces.19 • 0

0

Entering edit mode

I think people do have the same issue. I often get data from the sequencing core that has already been aligned to whatever version of the Ensembl genome they are using, and I then have to iterate through the AnnotationHub, testing for consistency in available Ensembl Gene IDs from a given Ensembl version and the data I have in hand until I can find the version that matches.

ADD REPLY • link 20 months ago James W. MacDonald 67k

0

Entering edit mode

Thank you!

I have been looking further into it and it looks like it's due to PAR as CellRanger uses GENCODE v32. The IDs are version controlled so I cannot map directly using annotationhub will try using UCS.

ADD REPLY • link 20 months ago sandra.garces.19 • 0

0

Entering edit mode

Hi, I have exactly the same problem here. May I ask how did you solve this at the end? I have around 200 genes that cannot be annotated by ens.hs.107 and I downloaded the reference (human (GRCh38)) from cellranger directly. Thank you!

ADD REPLY • link 12 months ago yw24418 • 0