Solution for pop-up problem with rsem-prepare-reference (RSEM can not recognize reference sequence name ATM...)

Question

[tsogurlu]

Hello everyone,

I was trying to use rsem-calculate-expression with the Aligned.toTranscriptome.out.bam file created by STAR. However, when I tried to run the code, I faced the error saying "RSEM can not recognize reference sequence name ATMG01375.1". I have searched for the solutions on the internet, and there was no explanation (At least, I couldn't). Then I decided to check the documents created by rsem-prepare-reference command manually. Then I have seen additional descriptions for this transcript in the all documents ("reference_name"). Here is the explanation:

>ATMG01330.1
GGGAGAGTGGTCAAAAGCGGCAGACTGTAAATCTGTTGAAGTTTTTCTACGTAGGTTCGAATCCTGCCTCTCCCA
>ATMG01340.1
ACCTACTTGACTCAGCGGTTAGAGTATCGCTTTCATACGGCGAGAGTCATTGGTTCAAATCCAATAGTAGGTA
>ATMG01375.1;Name=ATMG01375.1;id2=exon-id-trnH(GTG)-1;parent2=id-trnH(GTG)
GCGGATGTAGCCAAGTGGATCAAGGCAGTGGATTGTGAATCCACCATGCGCGGGTTCAATTCCCGTCGTTCGCC
>ATMG01380.1
AAACCGGGCACTACGGTGAGACGTGAAAACACCCGATCCCATTCCGACCTCGATATGTGGAATCGTCTTGCGCCATATGTACTGAGATTGTTCGGGAGACATGGTCCAAGCCCGGTGA

When it comes to how to solve the problem, I manually deleted these explanations. I do not know why it happened. I wanted to share this experience with the ones who faced this problem.

[Ram]

Is this in the transcripts.fa file? Can you show me the output to grep ATMG01375.1 gtf_file.gtf where gtf_file.gtf is the GTF file you used to prepare the reference?

[tsogurlu]

All of them (.transcript.fa, .idx.fa, .n2g.idx.fa, .seq,) included these explanations just for this transcript. Yes, it stems from the gtf file. Here it is:

ChrM    Araport11       tRNA    124603  124676  .       -       .       transcript_id "ATMG01375.1"; gene_id "ATMG01375";

ChrM    Araport11       exon    124603  124676  .       -       .       transcript_id "ATMG01375.1;Name=ATMG01375.1;id2=exon-id-trnH(GTG)-1;parent2=id-trnH(GTG)"; gene_id "ATMG01375:exon:1";

This gtf file is created this month (Araport11_GTF_genes_transposons.Jul2023.gtf)

EDIT: It seems the 2021 gtf file (Araport11_GFF3_genes_transposons.Mar92021_gffread.gtf) has also the same description: https://github.com/gpertea/gffread/issues/74 When I used Araport11_GTF_genes_transposons.Mar172021.gtf.gz, there was no problem.

[Ram]

This seems improperly formatted. Are there other entries with this sort of weird IDs?

[tsogurlu]

No, just this one: ATMG01375.1. At least, when I searched the word "Name", there was no other result.

[Ram]

OK then just edit that line and recreate the RSEM index.