How RSEM define gene

Question

[clingyun]

Hellow everyone,

When mapping reads to a de novo trinity reference using RSEM, I can get two results. One is the *genes.results, another is *isoforms.results. By checking the two files, it seems that the gene is combining of all the isoforms within a "gene" from trinity. However, I am not sure. I am thinking how RSEM define a gene? Does RSEM has it own methods to define a gene? I did not get answer from RSEM's algorithm paper and website.

Thanks for any suggetion

Chen Lingyun

Answer 1

[h.mon]

How did you run the analyses? Did you use Trinity's set of scripts? If so, Trinity probably automatically built a transcript to gene map, which will inform RSEM how to summarize genes from transcripts.

By checking the two files, it seems that the gene is combining of all the isoforms within a "gene" from trinity.

That's exactly what it is.

[clingyun]

Thanks!! Yes. I used the trinity to get the "Trinity.fasta", then used the supertranscript function in Trinity to get supertranscripts from "Trinity.fasta" The supertranscript was used for mapping using RSEM+STAR. Trinity built a relationship between isoforms and genes, which were provided to RSEM. I suspect that RSEM take the relationship directly. However, I did not see from RSEM's paper and manual.

[h.mon]

Could you show the command-lines of your complete pipeline?

[clingyun]

Thanks, sorry for late reply. The command is:

rsem-calculate-expression -p 8 --paired-end --star SRR5572144_1.fastq.trimmomatic.p SRR5572144_2.fastq.trimmomatic.p /home/clingyun/CHEN/database/Cquinoa_RSEM_STAR_genome/Cquinoa_RSEM_STAR /home/clingyun/CHEN/coexpression_network/SRR5572144/RSEM_STAR_results/SRR5572144

It can be abbreviated as: rsem-calculate-expression -p 8 --paired-end --star my_readfile1 my_readfile2 reference_dir output_dir