Issue with `seqnames` when adding gene/transcript annotation

[anamariaelek]

I get the following error when trying to add gene or transcript annotations to the coverage plot:

Error in stop_if_wrong_length("'seqnames'", ans_len) :                                                                                                                                                     
  'seqnames' must have the length of the object to construct (1) or                                  
  length 1   

I checked, and the seqnames of the gene/transcript annotation are overlapping the one shown in the plot, so I don't know what causes this error. It is produced by a function that creates new GRanges object under the hood, but I couldn't figure out why.

The gtf file I am loading looks like this:

GRanges object with 360449 ranges and 9 metadata columns:
              seqnames            ranges strand |   source       type     score
                 <Rle>         <IRanges>  <Rle> | <factor>   <factor> <numeric>
       [1] NC_064034.1     555751-623986      - |   Gnomon transcript        NA
       [2] NC_064034.1     555751-556233      - |   Gnomon exon              NA
       [3] NC_064034.1     557679-557838      - |   Gnomon exon              NA
       [4] NC_064034.1     558967-559103      - |   Gnomon exon              NA
       [5] NC_064034.1     559356-559885      - |   Gnomon exon              NA
       ...         ...               ...    ... .      ...        ...       ...
  [360445] NC_064048.1 13228100-13228198      - |   Gnomon       exon        NA
  [360446] NC_064048.1 13228331-13228935      - |   Gnomon       exon        NA
  [360447] NC_064048.1 13227468-13227632      - |   Gnomon       CDS         NA
  [360448] NC_064048.1 13228100-13228198      - |   Gnomon       CDS         NA
  [360449] NC_064048.1 13228331-13228558      - |   Gnomon       CDS         NA
               phase          transcript_id                gene_id    gene_name
           <integer>            <character>            <character>  <character>
       [1]      <NA> Nvec_vc1.1_XM_032374.. Nvec_vc1.1_LOC116613.. LOC116613690
       [2]      <NA> Nvec_vc1.1_XM_032374.. Nvec_vc1.1_LOC116613.. LOC116613690
       [3]      <NA> Nvec_vc1.1_XM_032374.. Nvec_vc1.1_LOC116613.. LOC116613690
       [4]      <NA> Nvec_vc1.1_XM_032374.. Nvec_vc1.1_LOC116613.. LOC116613690
       [5]      <NA> Nvec_vc1.1_XM_032374.. Nvec_vc1.1_LOC116613.. LOC116613690
       ...       ...                    ...                    ...          ...
  [360445]      <NA> Nvec_vc1.1_XM_048723.. Nvec_vc1.1_LOC125560.. LOC125560762
  [360446]      <NA> Nvec_vc1.1_XM_048723.. Nvec_vc1.1_LOC125560.. LOC125560762
  [360447]         0 Nvec_vc1.1_XM_048723.. Nvec_vc1.1_LOC125560.. LOC125560762
  [360448]         0 Nvec_vc1.1_XM_048723.. Nvec_vc1.1_LOC125560.. LOC125560762
  [360449]         0 Nvec_vc1.1_XM_048723.. Nvec_vc1.1_LOC125560.. LOC125560762
                  transcript_name      gene_type
                      <character>    <character>
       [1] Nvec_vc1.1_XM_032374.. protein_coding
       [2] Nvec_vc1.1_XM_032374.. protein_coding
       [3] Nvec_vc1.1_XM_032374.. protein_coding
       [4] Nvec_vc1.1_XM_032374.. protein_coding
       [5] Nvec_vc1.1_XM_032374.. protein_coding
       ...                    ...            ...
  [360445] Nvec_vc1.1_XM_048723.. protein_coding
  [360446] Nvec_vc1.1_XM_048723.. protein_coding
  [360447] Nvec_vc1.1_XM_048723.. protein_coding
  [360448] Nvec_vc1.1_XM_048723.. protein_coding
  [360449] Nvec_vc1.1_XM_048723.. protein_coding
  -------
  seqinfo: 31 sequences from an unspecified genome; no seqlengths

And the regions are:

       seqnames  start    end      score      Type
 1: NC_064034.1 554751 555575 0.00000000 muscle_1
 2: NC_064034.1 555576 555600 0.04792738 muscle_1
 3: NC_064034.1 555601 555650 0.00000000 muscle_1
 4: NC_064034.1 555651 555700 0.09585476 muscle_1
 5: NC_064034.1 555701 555725 0.00000000 muscle_1
 ---                                                        
26: NC_064034.1 556176 558125 0.00000000 muscle_2
27: NC_064034.1 558126 558175 0.07412679 muscle_2
28: NC_064034.1 558176 559350 0.00000000 muscle_2
29: NC_064034.1 559351 559400 0.07412679 muscle_2
30: NC_064034.1 559401 560885 0.00000000 muscle_2

I also attache the files to reproduce this, if it may help. Hope you could have an idea that will narrow down the source of the problem.

Thanks,
Anamaria

reproduce-issues-files.zip

[BioLaoXu]

I had the same problem, when I set the gene.name in the geom_transcript function it didn't report an error, you can try it.

[showteeth]

Sorry for the late response, I will look into it.

[jguppy]

Hi!

I'm running into the same issue when looking to add either the geom_gene or geom_transcript (ggcoverage v0.7.1). Unfortunately adding gene.name = "gene_a" into geom_transcript() doesn't help this time. Plotting the coverage plot itself works fine.

I had the issue whne using a gff3 file, and also when a custom grange object as per below.

Receiving this error;
Error in stop_if_wrong_length("'seqnames'", ans_len) : 'seqnames' must have the length of the object to construct (1) or length 1

As suggested previously I think it is coming one of the uses of GenomicRanges::GRanges(). I have found on GenomicRanges repo a similar error discussed in this issue Bioconductor/GenomicRanges#32

Any guidance you can provide here would be amazing, thanks for your time looking at this!

sample.meta = data.frame(SampleName = c("SRR1_chr19", "SRR1_chr19", "SRR1_chr19",
                                        "SRR1_chr19", "SRR1_chr19", "SRR1_chr19"),
                         Type = c("Male 1", "Male 2", "Male 3",
                                  "Female 1", "Female 2", "Female 3"),
                         Group = c("Male", "Male", "Male",
                                   "Female", "Female", "Female"))

track.df = LoadTrackFile(track.file = c("SRR1_chr19.bw", "SRR2_chr19.bw", "SRR3_chr19.bw",
                                        "SRR4_chr19.bw", "SRR5_chr19.bw", "SRR6_chr19.bw"),
                         format = "bw",
                         norm.method = "None",
                         meta.info = sample.meta) 
track.df$seqnames <- droplevels(track.df$seqnames)

#make a custom grange item for target transcript
df = data.frame(chr="chromosome19",
                start=20059287, end=20065261,
                strand="-", score="1",
                type = "exon",
                gene_name = "gene_a",
                gene_type = "exon",
                transcript_name = "gene_a")

gtf4.gr <- makeGRangesFromDataFrame(df, keep.extra.columns = T)

# plot coverage
basic.coverage = ggcoverage(data = track.df,
                            region = "chromosome19:20,058,000-20,066,000",
                            extend = 100,
                            facet.key = "Type",
                            group.key = "Group",
                            color = c("#f8766d","#f8766d","#f8766d",
                                      "#00bfc4","#00bfc4","#00bfc4"),
                            range.position = "out")+
  #ylim(c(0,155))+
  #geom_gene(gtf.gr=gtf.gr)+
  geom_transcript(gtf4.gr, gene.name = "gene_a")

basic.coverage

Running traceback() I get;

> traceback()
10: stop(wmsg(what, " must have the length of the object ", "to construct (", 
        ans_len, ") or length 1"))
9: stop_if_wrong_length("'seqnames'", ans_len)
8: new_GRanges("GRanges", seqnames = seqnames, ranges = ranges, 
       strand = strand, mcols = mcols, seqinfo = seqinfo)
7: GRanges(ans_seqnames, ans_ranges, strand = ans_strand, ans_mcols, 
       seqinfo = ans_seqinfo)
6: GenomicRanges::makeGRangesFromDataFrame(gene.tx.df.tx, ignore.strand = TRUE, 
       seqnames.field = c("seqnames"), start.field = "start", end.field = "end", 
       strand.field = "strand")
5: ggplot_add.transcript(object, p, objectname)
4: ggplot_add(object, p, objectname)
3: add_ggplot(e1, e2, e2name)
2: e1 %+% e2
1: `+.gg`(ggcoverage(data = track.df, region = "chromosome19:20,058,000-20,066,000", 
       extend = 100, facet.key = "Type", group.key = "Group", color = c("#f8766d", 
           "#f8766d", "#f8766d", "#00bfc4", "#00bfc4", "#00bfc4"), 
       range.position = "out"), geom_transcript(gtf4.gr, gene.name = "gene_a"))

[showteeth]

Hi @jguppy

Before I look into it, can you try the following:

• Upgrade ggcoverage with remotes::install_github("showteeth/ggcoverage"), the current version is 1.2.0
• Check the seqnames in SRR1_chr19.bw, the seqnames used to create GRanges should be the same. chromosome19 doesn't look like a common form.

[showteeth]

Hi all,

I figured this out!

This error 'seqnames' must have the length of the object to construct (1) or length 1 occurred when there were no gene value in gtf.gr, reference code.

Example:

# wrong gtf.gr
> gtf.gr$type %>% unique()
[1] exon CDS 
Levels: exon CDS
> basic.coverage + geom_gene(gtf.gr=gtf.gr)
Error in stop_if_wrong_length("'seqnames'", ans_len) : 
  'seqnames' must have the length of the object to construct (1) or length 1

Solution:

# add gene value to gtf.gr
gtf.gr$type = ifelse(gtf.gr$type=="exon", "gene", gtf.gr$type)
# now it works
basic.coverage + geom_gene(gtf.gr=gtf.gr)

[m-jahn]

@xf78 @anamariaelek @jguppy
Is this issue resolved for you?