vg rna index

Question

[Yumin]

Hi! I'm using vg rna for transcriptomic analyses following the wiki. I will list each of my steps and confusion below：

• First, I started with the final pangenome graph .gfa file obtained by the cactus-minigraph pipeline, which is in GFA 1.1 format. I converted it into GFA 1.0 format using gfa-wp: gfa-wp joint-full.gfa > joint-full.cov.gfa

The first two columns of the P line in joint-full.cov.gfa looks like:

P       COR_A#0#Chr10:0-69032057
P       COR_B#0#Chr10:107698-185749
P       COR_B#0#Chr10:198532-281912
P       COR_B#0#Chr10:291916-528385
P       COR_B#0#Chr10:759772-790564

• Then I used joint-full.cov.gfa for the GBWT construction:

  vg convert -p -t 40 joint-full.cov.gfa > joint-full.cov.pg
  vg gbwt -p -d ./ -o joint-full.cov.gbwt \
   --path-regex "(.*)#(.*)#(.*):(.*)" --path-fields _SHCF \
   -G joint-full.cov.gfa
  

• Finally, I used vg rna for Pantranscriptomes Construction:

  vg rna -p -t 40 -s ID -j -k 32 -u \
   -n gff/COR_A.gff -n gff/COR_B.gff \
   -n gff/COR_C.gff -n gff/COR_D.gff \
   -l joint-full.gbwt -b joint-full.rna.gbwt -f joint-full.rna.fa \
   -i joint-full.rna.txt joint-full.pg > joint-full.rna.pg
  

  The following error occurred:

  [vg rna] Parsing graph file ...
  [vg rna] Parsing haplotype GBWT index file ...
  [vg rna] Graph and GBWT index parsed in 25.7318 seconds, 18.4524 GB
  [vg rna] Adding transcript splice-junctions and exon boundaries to graph ...
  [transcriptome] ERROR: Chromomsome path "COR_A#0#Chr10" not found in graph or haplotypes index (line 1).
  

So, my question is:

• How do my steps need to be improved to successfully build a pan-transcriptome index using vg rna based on the cactus-minigraph gfa file.
• How to solve the chromosome path mismatch in the error message, how do I know what the chromosome path in the gbwt index looks like?

Answer 1

[colindaven]

Perhaps for your second question, vg gbwt is the way to go to list the chr paths and contig/sample names:

Step 7: Metadata (one input GBWT):
    -M, --metadata          print basic metadata
    -C, --contigs           print the number of contigs
    -H, --haplotypes        print the number of haplotypes
    -S, --samples           print the number of samples
    -L, --list-names        list contig/sample names (use with -C or -S)
    -T, --thread-names      list thread names

[Yumin]

Thank you for sharing your experience. I used vg gbwt -M -C -H -S -L -T joint-full.cov.gbwt to view the information in the gbwt index, which is as follows:

29379 paths with names, 12 samples with names, 12 haplotypes, 10 contigs with names
Chr1
Chr3
Chr2
Chr10
Chr7
Chr9
Chr6
Chr8
Chr4
Chr5
12
COR_A
COR_B
COR_C
COR_D
COR_E
COR_F
COR_G
COR_H
COR_I
COR_J
COR_K
COR_L
_thread_COR_A_Chr10_0_0
_thread_COR_B_Chr10_0_107698
_thread_COR_B_Chr10_0_198532
_thread_COR_B_Chr10_0_291916
_thread_COR_B_Chr10_0_759772
_thread_COR_B_Chr10_0_943427
_thread_COR_B_Chr10_0_1088850
_thread_COR_B_Chr10_0_1166426
_thread_COR_B_Chr10_0_1229383
_thread_COR_B_Chr10_0_1992992
_thread_COR_B_Chr10_0_2266779
_thread_COR_B_Chr10_0_2317201
_thread_COR_B_Chr10_0_2361082
_thread_COR_B_Chr10_0_2377269
_thread_COR_B_Chr10_0_2521740
_thread_COR_B_Chr10_0_2552854
...

The names of thread shown here are completely different from those in the gfa P line path, which confuses me as to how to make the chromosome paths match. Even if I modify the chromosome name of the gff file according to the above thread name, it still doesn't work.