ORFquant preprint clarifications

[nh13]

Reading your excellent ORFquant preprint, I had a few clarifications:

Ribo-seq reads were stripped of their adapters using cutadapt

• what set of adapters did you use?
• what settings for cutadapt did you use? (version too)

Reads aligning to rRNA, snoRNAs and tRNA sequences were removed with Bowtie2

• did you extract these features from the genome FASTA using the GTF to create a "to remove" reference FASTA?
• what settings did you use for bowtie? (version too)

Unaligned reads were then mapped with STAR45 using the hg38 genome and the GENCODE 25 annotation in GTF format

• what assembly did you use (ex. Ensembl release 100) and which contigs (only primary?)
• what settings did you use for STAR? Was it single pass, two pass?

[nh13]

I'd also be curious about your thoughts on ignoring 5' and 3' ends of the open reading frame from this paper:

For ribosome profiling, frequent read pile-ups are observed at the 50- and 30- ends of an open reading frame which are largely uninformative to a gene’s translational efficiency [4]. While these pile-ups can be indicative of true translation dynamics [56], current best-practices have more recently settled on ignoring these regions during read quantification and calculations of translation efficiency [3, 21]. By providing the --truncate argument during reference curation, the 50- and 30- ends of each coding region will be recursively trimmed until the speci- fied amounts are removed from coding space. A recursive strategy is required here as GTF file-formats split the CDS (coding sequence of gene) record into regions separated by introns. By default, 45 nt (nucleotides) will be trimmed from the 50-ends and 15 nt from the 30-ends recursively until the full length is removed from coding space, as is the current convention within the ribosome profiling field [3]. The resulting output file will then be used to process ribosome footprint libraries and their corresponding bulk RNA-Seq libraries. If generating a GTF file for use solely with general bulk RNA-Seq datasets, this file should not be truncated.

[lcalviell]

Hi, here a summary for the preprocessing I used.

Jurkat: This is cutadapt 1.8 with Python 3.4.3 Command line parameters: -m 18 -a CTGTAGGCACCATCAAT -o trim_cut.fastq -

HeLa, U2OS, HepG2 and K562: This is cutadapt 1.8 with Python 3.4.3 Command line parameters: -m 18 --discard-untrimmed -a TGGAATTCTCGGGTGCCAAGG -g GTTCAGAGTTCTACAGTCCGACGATC -o trim_cut.fastq -

HEK23: the fastq in the GEO accession should have been already trimmed, please let me know if I'm wrong. https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE73136

Bowtie 2 version 2.3.4.1 in -local alignment mode using default parameters. Human rRNA: https://www.ncbi.nlm.nih.gov/nuccore/555853 snoRNA, miRNA and tRNA were downloaded using the UCSC table browser. I can provide the fasta files if you need but I suspect there might be better ways to get those sequences.

STAR 2.6.0a with following parameters (single pass): STAR --genomeDir --readFilesIn --runThreadN 2 --outFilterMismatchNmax 2 --outFilterMultimapNmax 20 --chimScoreSeparation 10 --chimScoreMin 20 --chimSegmentMin 15 --outSAMattributes NH HI AS nM NM MD XS --outFilterIntronMotifs RemoveNoncanonicalUnannotated --alignSJoverhangMin 500 --outFileNamePrefix $nameexp"" --outReadsUnmapped Fastx --outSAMtype BAM SortedByCoordinate --quantMode GeneCounts --outWigType bedGraph --outWigNorm RPM --outSAMstrandField intronMotif --outSAMmultNmax 1 --outMultimapperOrder Random --limitBAMsortRAM 19000000000 --outFilterMismatchNoverLmax 0.04

hg38 genome with no scaffolds. GENCODE version25 annotation, which merges HAVANA and ENSEMBL annotation.

[lcalviell]

Not a fan of deleting regions. What I do often is to count reads over a certain mapping length (>25 nts) which is usually more than enough to remove artifacts IMHO. However there's no perfect way, and heuristics like that can be helpful in some datasets (e.g. high pileups at the +5 codons etc....).