4.8.2 smORF filter¶

Function¶

smorf_filter filters candidate ORFs generated by smorf_scanner.

It can filter ORFs by start codon, peptide length, ORF category, antisense status, primary/secondary priority, completeness, and Kozak-context evidence. Kozak scoring can use annotated ORFs, built-in models, an external PWM, or aligned Kozak-context sequences.

Input files¶

Input	Required	Description
Scanner message table	Yes, except with `--list-builtin-kozak`	`*.message.txt` generated by `smorf_scanner`.
Custom Kozak PWM	Required only for `--kozak-mode pwm`	External PWM matrix file.
Kozak sequence file	Required only for `--kozak-mode sequence`	Aligned Kozak-context sequences used to build a PWM.

Parameters¶

Parameter	Required	Description
`-i`, `--input`	Yes	Input ORF message table generated by `smorf_scanner`. Required unless `--list-builtin-kozak` is used.
`-o`, `--out-prefix`	Yes	Prefix of output files. Default: `ORF.filtered`.
`-s`, `--keep-start-codons`	No	Comma-separated start codons retained in the output. Default: `ATG,CTG,GTG,TTG`.
`-m`, `--min-aa`	No	Minimum ORF peptide length to keep. Default: `8`.
`-M`, `--max-aa`	No	Maximum ORF peptide length to keep. Default: `10000`.
`-c`, `--keep-categories`	No	Comma-separated ORF categories retained before filtering. Default: Multiple smORF categories.
`-r`, `--remove-categories`	No	Comma-separated ORF categories removed before filtering. Default: `same_frame_iORF,antisense_ORF`.
`-A`, `--keep-antisense`	No	Keep antisense ORFs. Without this option, sense ORFs are required. Default: `False`.
`-S`, `--keep-secondary`	No	Keep secondary or lower-priority overlapping ORFs. Without this option, primary ORFs are required. Default: `False`.
`-P`, `--keep-partial`	No	Keep incomplete ORFs without a complete start-stop structure. Default: `False`.
`-k`, `--kozak-mode`	No	Kozak model source: `none`, `annotated`, `builtin`, `pwm`, or `sequence`. Default: `annotated`.
`-b`, `--builtin-kozak`	No	Built-in Kozak PWM used when `--kozak-mode builtin`. Default: `plant`.
`-p`, `--kozak-pwm`	Conditional	Custom Kozak PWM matrix file. Required when `--kozak-mode pwm`.
`-q`, `--kozak-seq`	Conditional	Aligned Kozak sequence file. Required when `--kozak-mode sequence`.
`-C`, `--annotated-categories`	No	ORF categories used to build an annotated ORF Kozak PWM. Default: `annotated_ORF`.
`-n`, `--min-annotated-kozak`	No	Minimum number of annotated Kozak sequences required to build an annotated PWM. Default: `100`.
`-B`, `--fallback-builtin-kozak`	No	Built-in Kozak PWM used when annotated PWM construction fails. Default: `plant`.
`-F`, `--no-kozak-fallback`	No	Disable fallback to built-in PWM if annotated PWM construction fails. Default: `False`.
`-w`, `--min-kozak-pwm-score`	No	Minimum normalized Kozak PWM score required to pass. Default: `0.0`.
`-e`, `--export-kozak-pwm`	No	Export the loaded or constructed Kozak PWM to this file.
`-L`, `--list-builtin-kozak`	No	List available built-in Kozak PWM models and exit. Default: `False`.

Output files¶

Assuming -o smorf.filtered, the command writes:

Output	Description
`smorf.filtered.passed.message.txt`	ORFs that passed all filters. This is the main input for `smorf_evidence`.
`smorf.filtered.removed.message.txt`	ORFs removed by filtering.
`smorf.filtered.all.message.txt`	All input ORFs with added filtering/Kozak fields.
Custom PWM output	Optional file produced by `--export-kozak-pwm`.

Examples¶

Use the default annotated Kozak mode¶

smorf_filter \
  -i smorf.raw.message.txt \
  -o smorf.filtered \
  -s ATG,CTG,GTG,TTG \
  -m 8 \
  -M 1000 \
  -k annotated \
  -C annotated_ORF \
  -n 100

Use a built-in Kozak model¶

smorf_filter \
  -i smorf.raw.message.txt \
  -o smorf.filtered.plant \
  -k builtin \
  -b plant \
  -w 0.2

List built-in Kozak models¶

smorf_filter --list-builtin-kozak

Notes¶

--kozak-mode annotated builds a species-specific PWM from categories specified by --annotated-categories.
If there are not enough annotated ORFs to build a PWM, the command falls back to --fallback-builtin-kozak unless --no-kozak-fallback is used.
The default --remove-categories removes same_frame_iORF and antisense_ORF. Use --keep-antisense if antisense candidates should be retained.
The output file is named *.removed.message.txt, not *.failed.message.txt.