nf-core/rnastructurome

Path to comma-separated file containing information about the samples in the experiment.

The output directory where the results will be saved. You have to use absolute paths to storage on Cloud infrastructure.

Email address for completion summary.

MultiQC report title. Printed as page header, used for filename if not otherwise specified.

Path to transcript FASTA file.

Path to transcript annotation GTF file.

Ensembl FTP base URL used when auto-resolving transcript FASTA.

Ensembl release for transcript FASTA auto-resolution (current, latest, or release number like 114).

Fallback sample ID for rows missing sample_id.

Fallback probing method for rows missing method.

Fallback probing principle for rows missing principle.

Fallback organism for rows missing organism, typically a Latin binomial such as Homo sapiens.

Fallback chemical probing reagent for rows missing chemical (e.g. 1M7).

Fallback reverse transcriptase enzyme for rows missing RT_enzyme (e.g. M-MLV).

Fallback DMS reaction pH for rows missing pH.

Fallback UMI pattern for rows missing umi_pattern. Supplying a pattern enables umi_tools extract before cutadapt.

When true (default), a treated group with no exact sample_group+replicate untreated match falls back to the untreated sample sharing the same sample_group base token (portion before the first _) at the same replicate — e.g. MDA-MB-231_untreated_r1 can serve as control for MDA-MB-231_MTX_treated_r1. If a treated group still has no untreated after that, and exactly one untreated control exists anywhere else in the same reference, that single control is reused for it (with a warning) — e.g. one shared untreated backing several treated replicates. Set false to require exact matches; unmatched treated groups then proceed without an untreated control (scoring method 2 or 4).

Global fallback 5’ adapter sequence(s), comma-separated. If omitted and no per-sample adapter is provided, the pipeline defaults to AGATCGGAAGAGC.

Global fallback 3’ adapter sequence(s), comma-separated. If omitted and no per-sample adapter is provided, the pipeline defaults to AGATCGGAAGAGC.

If true, skip adapter trimming and apply quality/length trimming only.

Base quality threshold for 5’ trimming (used for MaP; RT-stop is forced to 0).

Base quality threshold for 3’ trimming.

Minimum read length retained after trimming.

Minimum adapter overlap required for trimming.

Trim terminal N bases.

Align against a transcript-level FASTA using Bowtie (RT-stop) / Bowtie2 (MaP). Default (false) downloads a genome FASTA and aligns with STAR.

Genome route (STAR) quantification method. Default (false) uses STAR –quantMode TranscriptomeSAM projected onto transcripts, tag-corrected with samtools calmd, counted with rf-count. Set true for the legacy rf-count-genome + rf-rctools extract path.

Maximum number of loci a read is allowed to map to in STAR (--outFilterMultimapNmax). Reads mapping to more loci than this are discarded. Increase when studying repetitive non-coding RNAs.

Report up to this number of alignments per read (-k).

Report all valid alignments (-a).

Trim this many bases from 5’ read end (--trim5).

Trim this many bases from 3’ read end (--trim3).

Optional path to a prebuilt Bowtie index (currently not consumed by workflow logic).

Bowtie v1 -n seed mismatch mode value.

Bowtie v1 -v mismatch mode value.

Bowtie v1 -m maximum reportable alignments per read before suppression.

Bowtie v1 --chunkmbs memory chunk size.

Bowtie2 alignment preset. Default is equivalent to --local -N 0 -D 20 -R 3 -L 20 -i S,1,0.50.

Bowtie2 --mp mismatch penalties as max,min.

Bowtie2 --dpad padding around DP table.

Bowtie2 --rdg read gap penalties as open,extend.

Bowtie2 --rfg reference gap penalties as open,extend.

Enable Bowtie2 local alignment mode (--local).

Bowtie2 --ma match bonus used with --local.

Allow dovetailing mates (--dovetail).

Skip SAMtools markdup deduplication (default: true). Position-based duplicate removal is not valid for chemical-probing libraries without UMIs, where reads sharing a 5’ start are independent RT events rather than PCR duplicates. Leave enabled; UMI libraries are still deduplicated via UMI-tools. Set false only if you have a specific reason.

STAR --sjdbOverhang for MaP alignment. Ideally readLength - 1; default 200 suits 201 bp reads.

Deprecated: rf-count plot generation (-g) is always enabled.

Library strandedness fallback for rf-count-genome when RSeQC inference is unavailable (e.g. viral/bacterial references without a usable BED annotation).

Minimum MAPQ score for rf-count-genome (--map-quality). Useful for filtering multi-mappers; 0 keeps all alignments.

Block size for per-chromosome memory allocation in rf-count-genome (-bs); increase for large genomes.

SAMtools working threads per processor instance in rf-count-genome (-wt); total threads = -p × -wt.

MaP only: discard reads with median Phred+33 base quality below this value (-eq).

MaP only: ignore all deletion events during mutation counting (-nd).

MaP only: ignore ambiguously aligned deletion events (-na). Auto: off for M-MLV, on for Group II Intron (e.g. TGIRT); leave unset to use the auto default.

MaP only: ignore insertion events during mutation counting (-ni). Auto: off for M-MLV, on for Group II Intron (e.g. TGIRT); leave unset to use the auto default.

MaP only: mark only the right-most base of a deletion as mutated (--right-deletion). Auto: on for M-MLV, off for Group II Intron (e.g. TGIRT); leave unset to use the auto default.

MaP only: discard mutations within N nt of each other (-dc N). Auto: 3 when RT_enzyme is a Group II Intron RT (e.g. TGIRT), off for M-MLV; leave unset to use the auto default.

MaP only: also evaluate quality of ±1 nt surrounding each mutation (-es).

Trim this many bases from the 5’ read end for rf-count (-t5).

Path to rf-count mask file (--mask-file).

Use only primary alignments in rf-count (--primary-only).

Discard reads with excessive 5’ or 3’ end clipping in rf-count / rf-count-genome (--discard-clipped-reads).

For paired-end samples, use only reads with both mates mapped (--paired-only).

For paired-end samples, use only properly paired reads (--properly-paired).

In mutation mode (-m), pre-sort paired-end reads by read name (--sort-by-read-name). No effect on single-end data.

In mutation mode (-m), discard reads shorter than this value (--discard-shorter).

In mutation mode (-m), minimum Phred+33 base quality for mutations (--min-quality).

In mutation mode (-m), collapse consecutive mutations/indels (--collapse-consecutive). Auto: on unless RT_enzyme is a Group II Intron RT (e.g. TGIRT); leave unset to use the auto default.

In mutation mode with collapsing enabled, max distance to collapse (--max-collapse-distance).

Remap normalized reactivities to the 0-1 range according to Zarringhalam et al. (--remap-reactivities).

Reactive bases used for normalization window selection (--reactive-bases), e.g. AC for DMS.

Normalization window size (--norm-window). Auto: 50 for DMS or RT-stop, unset otherwise; leave unset to use the auto default.

Normalization window offset (--window-offset).

Enable dynamic normalization window (--dynamic-window). Set to a positive integer to enable (value is passed to --norm-window); set to false or 0 to explicitly disable even for DMS (which enables it by default). Leave unset to use the per-method default.

Normalize each reactive base independently (--norm-independent).

Cross-experiment normalization via rf-normfactor (one factor set per reference, fed to rf-norm via -nf), putting reactivities on a common scale. Leave unset for auto (enabled only for a reference with more than one treated sample to cross-normalise); set true to force on, false to force off (per-sample box-plot).

rf-normfactor minimum coverage (-mc): bases below this coverage are excluded from the normalization factor calculation. Auto: 1000 for MaP, 50 for RT-stop; leave unset to use the auto default.

Override rf-norm scoring method (-sm). 1=Ding, 2=Rouskin, 3=Siegfried, 4=Zubradt.

Override rf-norm normalization method (-nm). 1=2-8%, 2=90% Winsorizing, 3=Box-plot, 4=Mitchell (MaP only).

Score raw reactivities without applying normalization (--raw).

Pseudocount used by Ding scoring (--pseudocount).

Set reactivities lower than untreated to zero for Ding or Siegfried scoring (--ignore-lower-than-untreated).

Discard transcripts with mean coverage below this threshold (--mean-coverage).

Discard transcripts with median coverage below this threshold (--median-coverage).

Positions with read coverage below this threshold are reported as NaN (--nan). Auto: 1000 for MaP, 50 for RT-stop. Set to 0 to disable NaN masking.

Genome route only: after rf-rctools extract, keep transcripts with at least one position at this coverage or higher before rf-norm. Set to 0 to keep the full annotation RC.

Deprecated: rf-norm plot generation (--img) is always enabled.

Generate ViennaRNA RNAplot structure diagrams via rf-fold (-g). Disabled by default; slow on large transcriptomes. Use --r2dt for template-based diagrams instead.

Path to a reference .db structure file for rf-jackknife normalisation assessment. When provided, rf-jackknife runs between rf-norm and rf-fold and the output CSV reports optimal slope/intercept values. When omitted, rf-fold runs directly using whatever slope/intercept params are set.

Comma-separated slope range to search in rf-jackknife (-sl), e.g. 0,5.

Comma-separated intercept range to search in rf-jackknife (-in), e.g. -3,0.

Step size for slope grid search in rf-jackknife (-ss).

Step size for intercept grid search in rf-jackknife (-is).

Use modified FMI (Lan et al. 2022) instead of standard FMI in rf-jackknife (-m).

Use relaxed FMI criteria (Deigan et al. 2009) in rf-jackknife (-x).

Retain lonely base-pairs (1 bp helices) in the reference structure before comparison (-kl).

Retain pseudoknotted base-pairs in the reference structure during FMI comparison (-kp). Disabled by default because rf-fold cannot predict pseudoknots, so including them systematically lowers FMI.

Only use transcripts present across all replicates in rf-jackknife (-oc).

Generate R heatmap of grid-search results in rf-jackknife (-g).

Additional rf-fold parameters passed inside rf-jackknife (-rp), e.g. -md 500.

Pool XMLs from all sample groups into a single rf-jackknife run (output: jackknife/all_groups/) instead of running one jackknife per group. Default: true.

Stop the pipeline after rf-jackknife completes, skipping rf-fold, structure visualisation, browser track generation, and RDAT output. Useful for calibration runs where you only want jackknife statistics.

GTF feature type to extract with rf-rctools (-f); override for non-standard annotations.

GTF attribute used as the output RC entry ID by rf-rctools (-b).

Path to a .db file of known RNA secondary structures. Enables rf-eval when provided.

Cutoff for classifying a base as highly-reactive in rf-eval unpaired-coefficient calculation (-c).

Exclude terminal base-pairs from rf-eval calculations (-it).

Treat terminal base-pairs as unpaired in rf-eval (-tu).

Retain pseudoknotted base-pairs in rf-eval (-kp).

Retain lonely/isolated base-pairs in rf-eval (-kl).

Generate R metric plots in rf-eval (-g).

Run rf-structextract after rf-fold to extract high-confidence, low-reactivity / low-Shannon structural motifs from the folded structures.

rf-structextract window size in nt for median reactivity/Shannon calculations (-w).

rf-structextract skips low-reactivity/low-Shannon evaluation for transcripts below this length (-ml).

rf-structextract windows with less than this fraction of bases covered are set to NaN (-mv).

rf-structextract minimum fraction of bases whose Shannon and reactivity are below the transcript median (-mb).

rf-structextract discards elements with less than this fraction of paired bases (-mp).

rf-structextract discards structure elements shorter than this (-mm).

rf-structextract discards structure elements longer than this (-xm). Unset for no limit.

rf-structextract discards elements with a loop larger than this (-xl). Unset for no limit.

rf-structextract skips low-reactivity evaluation (-ir). When false (default), reactivity is evaluated so only regions with probing support are extracted.

rf-structextract skips low-Shannon evaluation (-is). When false, the Shannon-entropy test is applied so only high-confidence regions are extracted.

rf-structextract only reports elements encompassing multiway junctions (-mo).

rf-structextract reports each extracted element in a separate file (-opf).

rf-structextract only reports elements with a free energy significantly lower than expected by chance (-ee).

rf-structextract p-value threshold for energy significance (-v); requires --structextract_eval_energy.

rf-structextract number of sequence shufflings for energy evaluation (-ns).

rf-structextract preserves dinucleotide frequencies when shuffling for energy evaluation (-ds).

Render an SVG diagram per extracted rf-structextract motif via ViennaRNA RNAplot.

Run rf-correlate to compute pairwise reactivity-profile correlations between replicates of each sample group (a replicate-reproducibility QC surfaced in MultiQC). Only runs for sample groups with more than one replicate; a no-op otherwise.

Cap reactivities to this value before Pearson correlation in rf-correlate (--cap-react). Not applied to the Spearman run, which is rank-based.

rf-correlate minimum number of values to calculate a correlation (-m); a value between 0 and 1 is interpreted as a fraction of transcript length. Unset uses the tool default (off).

rf-correlate ignores sequence differences (e.g. SNVs) between compared transcripts (-I).

Generate the rf-correlate correlation heatmap PDF (-g; requires R).

Write CT format structures with rf-fold (-ct). Disabled by default to keep dot-bracket output.

Enable windowed MFE folding in rf-fold with this window size (-w -fw N). Set null/unset to fold the whole transcript instead.

Partition-function window size in rf-fold (-pw N).

Fold without reactivity constraints in rf-fold (-u).

Pass ViennaRNA no-lonely-pairs mode to rf-fold (-vnlp).

Use ViennaRNA hard constraints with rf-fold (-vc).

Maximal base-pair span for ViennaRNA in rf-fold (-vmd).

Minimal base-pair span for ViennaRNA in rf-fold (-vms).

Only fold transcripts covered in at least this number of XML experiments (-oc).

Constraint file for allowed base-pairing positions (-fc).

Constraint file for required unpaired positions (-uc).

Generate dot plots from rf-fold (-d).

Compute and report Shannon entropy in rf-fold (-sh).

Slope for reactivity-to-folding-constraint conversion in rf-fold (-sl). Auto by chemical: DMS 4.6, NAI 2.2, 2A3 1, other/unset 1.8. Overridden by jackknife calibration when available.

Intercept for reactivity-to-folding-constraint conversion in rf-fold (-in). Auto by chemical: DMS -2, NAI -0.8, 2A3 -0.4, other/unset -0.6. Overridden by jackknife calibration when available.

Path to ViennaRNA RNAplot binary for SVG structure plots in rf-fold (-vrp).

Generate template-based 2D RNA structure diagrams using R2DT, coloured by normalised SHAPE/chemical-probing reactivity.

Container image used for the local R2DT module. The default is a digest pin of the rnacentral/r2dt :latest tag. An earlier pin shipped the ribovision-ssu/lsu, rnasep and tmrna covariance models without their cmfetch SSI indexes; because the data dir is read-only under Singularity, R2DT could not build them at runtime and those templates (e.g. LSU rRNA) failed to draw.

Comma-separated GTF biotypes sent to R2DT. R2DT only has templates for structured ncRNA classes; transcripts of other biotypes (mRNA, lncRNA, …) are excluded so R2DT never force-fits them to the wrong template, and are drawn by ViennaRNA instead. Matched case-insensitively against transcript/gene biotype.

Report positions with zero reactivity in the WIG/BigWig output (-z). Enabled by default so genome browsers receive complete tracks without gaps.

Restrict WIG output to specific bases using an IUPAC code or combination (e.g. AC for DMS probing) (-kb). Default: all bases (N).

Container image used for local RNAframework modules.

Path to R executable used by rf-count, rf-norm, and rf-fold plotting.

SLURM partition / queue name passed via --partition.

SLURM account / allocation passed via --account.

Extra sbatch flags appended verbatim to process.clusterOptions.

Custom MultiQC yaml file containing HTML including a methods description.

Display the help message.

Display the full detailed help message.

Display hidden parameters in the help message (only works when –help or –help_full are provided).