I am analysing Ribo-Seq data (Illumina single end ~25 bp reads) using the Xtail pipeline. I also have RNA-Seq data sequenced using the same samples and sequencing chemistry (Illumina single end ~25 bp reads). During the bioinformatic analyses, we figured out that in around 48% of genes the Ribo-seq count is more than the RNA-Seq count for both treatment and control data. Strangely, there are also few genes (~3%) where we have ribo-seq reads but zero RNA-seq read count. Could anyone let me know if this is usual in such data analysis? However biologically this is somehow unbelievable to have Ribo-Seq of genes which are not transcribed (no RNA-Seq).
We can imagine that some RNAs are in low absolute concentration in your RNA-seq total sample, but that they are translated with high efficiency and end up enriched in the Ribo-seq sample. Because they are enriched, they would account for higher proportions of the library, so I don't find your results so surprising.
By the way, you can not say that a gene is not transcribed because you have 0 reads in RNA-seq. What you can say is that the expression of that genes (the resultance of transcription and degradation) falls below the level of detection of your study. If you had sequenced more reads, it is possible that you would find >0 reads for that gene.
Still, I would argue seeing Ribo-seq reads mapped to the genes without RNA-seq readouts doesn't only mean you have a higher efficiency of translation for some very low abundant genes in mRNA level. It more means you maybe have some technical issues with RNA-seq design and it might be that you need to increase the depth of you RNA-seq experiment. Especially you see that 48% genes have more Ribo-seq reads than RNA-seq. I would also check the 3nt periodicity for the Ribo-seq reads over annotated CDS genes to see if your Ribo-seq reads are doing well in detecting annotated CDS.
