How to use Ribo-seq ORFs

By default, TE is estimated based on RPFs and RNA abundance for type=CDS (3rd column in GTF file). If Rp-Bp has been used for Ribo-seq ORFs discovery, this set of Ribo-seq ORFs can be used to generate a new GTF file, and TE can be estimated for Ribo-seq ORFs, instead of genes (CDS or exon). This also works with a de novo assembly, see e.g. More about de novo ORF discovery.

Hint

The same is true for DE, i.e. abundance is calculated for type=CDS by default, using gene features. But you can find regulated Ribo-seq ORFs across conditions, just follow the description below.

Prepare input for abundance estimation

You first need to locate the output of summarize-rpbp-predictions, in particular the BED12+ file <project_name>[.note][-unique][.filtered].predicted-orfs.bed.gz, see Summarizing the Rp-Bp predictions. This file contains the combined predicted translation events, or Ribo-seq ORFs, from all samples and replicates, with additional columns. To prepare a GTF file with the Ribo-seq ORFs

get-gtf-from-predictions [options] bed gtf

where bed is the path to the BED12+ Ribo-seq ORFs, as described above, and gtf is the path to the GTF file to be created, e.g. <project_name>[.note][-unique][.filtered].predicted-orfs.gtf. This file contain all the Ribo-seq ORFs, with CDS and exon features, and additional attributes, such as ORF type. For all options, consult the API for get-gtf-from-predictions.

You are now ready to call

run-htseq-workflow ribo [options] --htseq-options "--idattr orf_id" "--additional-attr orf_type" "--additional-attr gene_name" --gtf HTSEQ_GTF config

The important options are

  • --htseq-options "--idattr orf_id", this will indicate htseq-count to use the ORF id as feature for counting. You can add additional options, such as "--additional-attr orf_type", etc.

  • --gtf HTSEQ_GTF, where HTSEQ_GTF is the newly created GTF file <project_name>[.note][-unique][.filtered].predicted-orfs.gtf, this will indicate the program to use the Ribo-seq ORFs for abundance estimation, and not the existing GTF file.

Estimate TE or DE using Ribo-seq ORFs

Finally, you need to check the htseq-count output tables, to determine which column contains which attributes. Currently, htseq-count outputs "--additional-attr in the order as they are given, but this behaviour is not guaranteed. In particular if using multiple additional attributes, you need to make sure the right column is used, and they can be specified with -orfCol COLUMN_NUMBER and -symbolCol COLUMN, where orfCol is the ORF type column, if used, and symbolCol is the gene_name column. As given above, this would be --orfCol 2 --symbolCol 3.

run-tea -orfCol 2 --symbolCol 3 --delim TAB config

or

run-dea --orfCol 2 --symbolCol 3 -delim TAB config

Note

By default, DESeq2 does not specify the field separator for read.table, i.e. is uses the default white space. If some GTF attributes are missing, e.g. gene_name with a de novo assembly for novel Ribo-seq ORFs, htseq-count will omit the attribute, and the count file will not be read correctly. In this case, you should specify the field separator using --delim TAB.