Hi, <!channel> I'm new to riboseq data analysis. I have a single-end sequencing of the 2 samples of riboseq and my fragment size is about 75bp mean fragment size is 45 bp. I tried using the existing pipeline, but 99% of reads were not aligned because of two short regions, so I did the STAR aligner separately using the --sjdbOverhang 48 functions in index then aligned the reads 78% of the sequence is aligned to it. can any one help me with it or guide me in the down stream analysis

Is that 75bp raw and mean 45bp post adapter trimming? 45bp is still quite long for a RPF so it may warrant further inspection. May I ask why you chose 48 for your overhang? I believe this is 'supposed' to be max read length minus 1. 78% mapped is not bad but I always feel it is best to have a good grasp of these things And finally, what are you looking to investigate with this analysis? If you have a question in mind it might be easier to advise

@Jack Tierney Thanks. First I tried it with the 74 bp for the overhang; my alignment was only at 4. Then I checked the mean distribution of the fragments in the Fastq report I found that the majority of my fragments were in the range of 45 to 50, so I kept in the cut of 48 then my aligned score increased to 78%. i just want to check which translatome is upregulated in my treated sample, compared to the control. I have the RNA level expression level; now I want to check the translatome level. that's why I did riboSeq

hello, the thing is it's hard to tell what is going on without knowing whether your reads have been trimmed correctly or not. for monosome data footprint sizes usually peak in the range of 28-30sth nts depending on RNAse used for fragmentation etc . So if you aimed to profile monosomes, you'd expect your trimmed read lengths to be around that range. So in your case the longer length range can either be explained by : (a) adaptors not trimmed properly or (b) sth went wrong during the experiment - incomplete digestion, or wrong size selection or massive rRNA contamination/other contamination or any combination of (hard to say without knowing the protocol). For data analyses you need to make sure that: 1. all adaptor are trimmed completely and reads are clean because reads are mapped end-to-end in riboseq and 2. removal of rRNA is enabled (

--remove_ribo_rna

). sharing your fastqc after trimming and enabling rRNA removal might make a difference. these are some riboseq-specific considerations; if too few reads map for your RNA-seq data as well, it could be sth else. Generally, I would advise against tinkering the STAR settings too much - defaults are pretty optimised - you might end up being overly lenient and get false-positve mappings..

@Ira Iosub Thanks for your explanation. I have shared my FastQC report with you for you to take a look at. The sequence length distribution of he library is constructed for the 150 BP, and the majority of the reads are at 45-49. hence I'm playing around with the STAR settings to make sure my alignment rate is higher.

thanks! can you please clarify if you pre-mapped your reads to rRNA or not? this is crucial before proceeding to alignment with STAR; while technically you may be right in needing to reduce the

sjdbOverhang

param from the default of 100 for riboseq, where most reads are expected to be short, if rRNA reads havent been removed you may be seeing artifacts by tuning this paramater. Using

--sjdbOverhang 100

with rRNA-contaminated Ribo-seq can lead to reads being "too short" and unmapped. Lowering it allows them to map — but possibly spuriously. My advice is to make sure you are not in this scenario, especially given your length sizes are longer than what's expected of genuine RPFs.

Yes, i have pre-mapped my data with rRNA seq.