4.8 smORF analysis
Function
The smORF analysis module provides a complete workflow for scanning candidate small open reading frames (smORFs), filtering and clustering candidates with sequence and Kozak-context criteria, evaluating Ribo-seq translation evidence, and quantifying P-site density for reliable smORFs.
The workflow is designed for transcript-centric smORF discovery. Candidate ORFs are first generated from a genome FASTA and genePred annotation, then filtered and evaluated using P-site density tracks from Ribo-seq data.
Workflow
smorf_scanner → smorf_cluster → smorf_evidence → smorf_quant
| Step |
Command |
Main purpose |
| 1 |
smorf_scanner |
Scan transcript-centric candidate ORFs from genome FASTA and genePred annotation. |
| 2 |
smorf_cluster |
Filter scanned ORFs by sequence and Kozak-context criteria, then cluster them into non-redundant ORF families. |
| 3 |
smorf_evidence |
Evaluate family-aware Ribo-seq translation evidence and classify reliable smORFs. |
| 4 |
smorf_quant |
Quantify raw P-site density counts for reliable smORFs from per-sample density tracks. |
| Input |
Used by |
Description |
| Genome FASTA |
smorf_scanner |
Genome sequence used to reconstruct transcript sequences. |
| genePred annotation |
smorf_scanner, optional for smorf_evidence |
Transcript annotation. Scanner uses it to define transcript structures; evidence uses it to recover ORF exon blocks when needed. |
| Scanner message table |
smorf_cluster |
Candidate ORF table generated by smorf_scanner. |
| Clustered ORF table |
smorf_evidence |
Clustered family table generated by smorf_cluster. |
| P-site density files |
smorf_evidence |
Strand-specific or unstranded bedGraph/WIG files containing Ribo-seq P-site density. |
| Reliable smORF genePred |
smorf_quant |
Reliable smORF genePred generated by smorf_evidence. |
| Density list |
smorf_quant |
Design table pointing to the per-sample P-site density tracks. |
Main evidence levels
| Evidence |
Meaning |
| ORF sequence |
Start codon, stop codon, ORF length, strand, and category. |
| Kozak context |
Start-codon context scored by annotated, built-in, PWM, or sequence-derived Kozak models. |
| Ribo-seq signal |
Total RPF signal, covered nucleotides/codons, and coverage ratio. |
| Periodicity |
Frame-specific signal distribution, especially frame-0 enrichment. |
| Start site |
Resolved start site, leading-window support, and noncanonical extension. |
| Coverage shape |
Uniform, disperse, or skewed RPF distribution across the ORF. |
| Multi-sample support |
Reproducibility and support level across multiple Ribo-seq samples. |
| Quantification |
Raw P-site density count matrix for reliable smORFs across samples. |
Notes
smorf_scanner produces many candidate ORFs. Scanner output alone should not be treated as evidence of translation.smorf_cluster filters and clusters candidates according to sequence features and Kozak context, but it does not use Ribo-seq evidence.smorf_evidence requires a density list pointing to the P-site density tracks of every sample.smorf_quant is most useful when multiple Ribo-seq samples or replicates are available.- Strand-specific P-site density files can be generated with
rpf_Bam2bw and then supplied to smorf_evidence as bedGraph or WIG tracks.