Written guide to selected ORF and translation results. Bioinformatics → Sequence Analysis → ORF Detection (bioinformatics/Orf Detection, catalog ID TB01-APP04) searches a nucleotide sequence under explicit start, stop and reading-frame rules. This guide uses saved human ACTB results and the current accepted app’s source contract. It adds no new ORF scan, translation run, recording or whole-app acceptance.
1. Load one identified nucleotide record
The example uses transcript NM_001101.5 and its saved deposited protein reference NP_001092.1. The reviewed transcript FASTA contains 1,812 nucleotides; the reference protein contains 375 residues. Preserve accession versions and the exact input hash from the guide evidence inventory. A gene name such as ACTB is not a complete input identity.
In the launcher, choose Browse..., select the single-record nucleotide FASTA and choose Load. Then open Sequence Analysis → ORF Detection. Use the transcript as input and keep the protein FASTA separately for comparison. The app uses the first sequence record; a multi-record FASTA is not an instruction to batch every record.
The current input contract removes whitespace, uppercases bases and normalizes U to T. It accepts A/C/G/T/U/N and requires at least 30 nucleotides. Other ambiguity symbols are refused. N is retained with a warning; a start or stop spanning N cannot define an ORF, and an ambiguous translated codon becomes X. Inspect ORF Detection Contract QC, including sequence length and warnings, before interpreting a candidate. Missing input is not replaced by a synthetic sequence.
2. Keep the scan policy and selected routes explicit
The current app initially selects all 20 routes. For the saved translation example, retain only ORF Finder (6 frames), Longest ORF, Translation and Coding Potential (unavailable), then choose RUN. The unavailable route is included to make its boundary visible. These receipts do not validate all the other routes.
The finder scans three forward frames and three reverse-complement frames. Within each frame it selects the first ATG in a stop-delimited interval, then resumes after the stop. It does not enumerate every nested ATG. The minimum interval is 30 nucleotides; a sequence-end interval without a stop can be retained as partial. The app uses NCBI standard code 1 with an ATG-only start policy. This is not a configurable choice of every NCBI genetic code or every possible initiation codon.
The saved run found 24 candidate ORFs and translated the 15 longest nucleotide intervals, leaving nine untranslated. Ties retain forward +1/+2/+3, then reverse −1/−2/−3 scan order. Fifteen exported proteins means fifteen selected candidates, not fifteen confirmed genes. A new input or different policy can change both counts and ordering.
3. Distinguish scanned-strand coordinates from input coordinates
Intervals are zero-based and end-exclusive: [start, end) contains end − start nucleotides. For a negative frame, start and end refer to the reverse complement. Use input_start and input_end to compare positions on the uploaded transcript. For sequence length N, reverse interval [a, b) maps to [N − b, N − a).
| Saved candidate | Scanned strand → uploaded transcript | Length and interpretation |
|---|---|---|
| ORF_001, frame +1 | [84, 1212) → [84, 1212). The same positions are 85–1212 in one-based inclusive notation. | 1,128 nt, including terminal TAG. 1128 / 3 − 1 = 375 residues. Its complete saved protein agrees with the saved deposited ACTB reference. |
| ORF_002, frame −2 | [622, 1012) → [1812 − 1012, 1812 − 622) = [800, 1190). | 390 nt, including a stop; 129 translated residues. This is a candidate under the scan policy, not a deposited ACTB coding annotation. |
Nucleotide intervals include a terminal stop when present. Protein sequences exclude the stop and incomplete trailing codons. A nucleotide length divided by three is therefore not always the exported protein length. Retain the stop-codon and trailing-base fields rather than inferring completeness from length alone.
4. Read complete proteins beyond the table preview
The Translation table’s protein_50aa field is a preview. Select the named sequence result, such as Translation ORF_001, to inspect the complete protein; use Save Sequence to save that sequence as FASTA. In the saved example, the full ORF_001 file contains all 375 residues. The separate Save Sequence output agrees with it and with the saved deposited reference.
The est_mw_da column uses a rough 110 Da per residue rule. For 375 residues this is 41,250 Da. It is a length-based estimate, not a sequence-specific molecular mass or a measurement. The guide checks full FASTA lengths, preview correspondence and saved coordinate fields; it does not rerun biological translation.
Coding Potential supplies no probability or coding/noncoding verdict because a validated predictive model is unavailable. The historical native run records one unavailable method. Its CSV uses the generic status not_uploaded; read the accompanying reason, which identifies the missing predictor. That label does not mean the ACTB transcript failed to load. Preserve the unavailable result rather than replacing it with a score or treating it as successful prediction.
5. Read maps, overlaps and density with their reporting limits
A separate saved run selected ORF Visualization, ORF Overlap and ORF Density. It is a different receipt and export bundle. The map table retains all 24 uploaded-sequence intervals: 14 forward and 10 reverse. Compare both strands on that common input coordinate axis.
The overlap result counts 35 overlapping candidate pairs, reports the largest 25 in its table and displays 20 in its plot. Ten pairs are omitted from the table. Complete export of that table does not remove the method’s reporting limit. These are candidate-interval intersections, not proof of overlapping expressed genes.
Density uses 1,000-nucleotide windows at a 500-nucleotide step. Final windows may be shorter, and an ORF intersecting multiple windows is counted in each. For [1000, 1812), 13 intersections over 812 nt give about 16.00985 per kb; for [1500, 1812), seven over 312 nt give about 22.43590 per kb. This measures interval intersections per kilobase, not covered-base fraction or confirmed gene density. The displayed high-density threshold is descriptive, not a calibrated biological prediction.
6. Export and reopen the named results
- Retain the input FASTA, accession/version, selection policy, selected routes and QC/availability messages. Choose an empty destination directory so this export is easy to distinguish from earlier runs.
- Use the viewer’s Export All. The saved translation bundle contains 26 listed artifacts: four PNGs, five CSV tables, 15 complete protein FASTAs, one summary and one JSON analysis report. Its
artifact_manifest.jsonis separate, making 27 files in that directory. - Read the manifest’s titles, types, byte counts and SHA-256 digests. Identify proteins by their named sequence results, not directory order. The current exporter adds filename suffixes when names already exist; do not assume every run has the same filenames.
- Reopen each FASTA and count sequence residues without its header or line breaks. Compare those lengths with
protein_length_aa; compare only the first 50 residues with the preview field. Read all CSV rows, including coordinate policy, omissions and unavailable reasons. - Keep the geometry bundle separate: its nine listed artifacts are three PNGs, four CSVs, a summary and an analysis report, plus its own manifest. The report lists result titles and types; it does not replace the CSVs or FASTAs. Neither bundle promises a reloadable scientific session.
- Preserve the manifest itself with the other files and open plots outside the viewer. A checksum verifies bytes, not scientific validity. Review provenance and identifiers before sharing; historical saved artifacts may contain local paths.
Evidence, versions and further learning
Read the exact source and saved-export inventory. Five inspected current source files match the accepted Bioinformatics payload. The saved translation and geometry runs have different, earlier app/base/viewer hashes; they remain historical evidence. An even earlier receipt described a preview-only translation export and does not supersede the later complete-FASTA bundle. This guide adds saved-file integrity and arithmetic checks, not current scientific execution or a new movie.
The deposited ACTB comparison supports one identified transcript/protein. The other candidate proteins have the recorded algorithm-comparison scope, not deposited biological annotation. The original GenBank accession was not freshly re-fetched for this guide. Current source binding does not establish all 20 routes, hosted execution, real Windows/macOS usability or full-product acceptance.
Continue to RNA-seq normalization and exported precision for count-matrix inputs, or browse the written app guides for other workflows.