Output Files

Most ChromoSort commands write both sequence files and audit tables. FASTA outputs are intended for downstream assembly workflows; TSV, HTML, plot, and text reports are intended to make every keep, reject, split, cut, scaffold, or graph-fill decision inspectable.

AGP And Provenance Sidecars

Every ChromoSort command that writes a modified FASTA also writes a fresh AGP 2.1 map for that FASTA, a component-provenance TSV mirroring the AGP rows, and a submission-oriented checklist. Downstream commands should compose a new AGP for their own output rather than appending rows to an upstream AGP, because each FASTA-changing step creates a new object coordinate system.

The AGP file is the structural map: output objects, component slices, inserted N gaps, orientations, and part numbers. The component TSV and command-specific reports carry the richer audit trail: whether a row was unchanged, algorithmic, reviewed, manual, graph-filled, or a fallback gap. For submission workflows, keep the final FASTA, final AGP, component TSV, command reports, and *.submission_checklist.tsv together.

Output Index

Table 1. ChromoSort command output map. Use this as a quick index to the command-specific output sections below.

chromo sort Outputs

Use chromo sort outputs to inspect contig assignment, duplicate-overlap filtering, reference coverage, and the final reference-ordered FASTA.

Table 2. Example contig_assignments.tsv rows. Selected columns from the synthetic fixture show common status classes.

Table 3. Example chromosome_summary.tsv rows. The summary table groups retained contigs by reference.

Listing 1. Example ordered FASTA headers. Headers retain assignment and orientation metadata for audit.

>chr1_contigA original=contigA ref=chr1 ref_start=1 ref_end=80 orientation=+ reverse_complemented=no query_cov=1.0000 avg_identity=99.000
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
>chr2_contigC original=contigC ref=chr2 ref_start=10 ref_end=69 orientation=- reverse_complemented=yes query_cov=1.0000 avg_identity=97.000
CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

chromo clean Outputs

Use chromo clean outputs to inspect the full conservative cleanup decision: which raw contigs were discarded by sort logic, which retained contigs were selected for fix planning, which split plans were accepted or rejected, and what records were emitted into the cleaned FASTA.

Example clean_contigs.tsv rows. Selected columns show a discarded raw contig, two emitted split pieces, and one retained unsplit contig.

chromo fix Outputs

Use chromo fix outputs to inspect which requested or automatically detected contigs were split, which were copied unchanged, and why each breakpoint plan was accepted or rejected.

Table 4. Example split report rows. Selected columns from a sensitive-mode fixture show how one source contig can become multiple reference-labeled pieces.

Listing 2. Example fixed FASTA records. Split-piece headers carry the original contig, slice interval, alignment interval, and orientation.

>chrom02-contig_04-a original=contig_04 ref=chrom02 slice=1-20 alignment=1-20 orientation=+ reverse_complemented=no avg_identity=100.000
AAAAAAAAAAAAAAAAAAAA
>chrom07-contig_04-b original=contig_04 ref=chrom07 slice=21-40 alignment=21-40 orientation=+ reverse_complemented=no avg_identity=100.000
CCCCCCCCCCCCCCCCCCCC

chromo eval Outputs

Use chromo eval outputs when a task should be reviewed in a spreadsheet or in task-specific chromo manual dashboards before a sequence-changing command runs.

These tables use schema chromosort-review-event-v1. Rejected or deleted rows do not apply sequence changes. Task-specific manual dashboards load the same tables with --review-table and render their evidence fields as event panels.

chromo cut Outputs

Use chromo cut outputs when exact reviewed breakpoints should be applied without invoking the alignment-based chromo fix planner.

Table 5. Example cut report rows. Cutting contigA after bases 20 and 50 emits three adjacent slices and records the complete cut set on each row.

Listing 3. Example cut FASTA records. Requested contigs are replaced by cut pieces; unrequested contigs are copied through unchanged.

>contigA_cut001
AAAAAAAAAAAAAAAAAAAA
>contigA_cut002
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
>contigA_cut003
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
>contigB
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT

chromo manual Outputs

Use chromo manual outputs when reviewed browser curation should be exported and reproduced from the command line.

Manual scaffold-mode gaps default to AGP contig/no/na, which records an unlinked N gap without claiming evidence. When a reviewed manual scaffold join does have support, set --agp-gap-type, --agp-linkage, and --agp-linkage-evidence explicitly during chromo manual apply.

Table 6. Example chromo manual apply report rows. The command-line apply step writes piece-level audit information for a browser-reviewed recipe.

Listing 4. Example manual-apply FASTA records. Browser edits become ordinary FASTA records with the original source contig, slice interval, strand, and scaffold label in the header.

>contigA_manual001 original=contigA slice=1-80 strand=+ scaffold=chr1
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
>contigB_manual002 original=contigB slice=1-40 strand=+ scaffold=chr1
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT

chromo gafprep Outputs

Use chromo gafprep outputs when you want targeted read-to-graph evidence without aligning every read to the full graph. The command writes inputs for GraphAligner and audit tables explaining why each read was selected.

If summary files include GAF evidence limited by graph sanitization near target, treat GAF results for that region as limited review evidence.

chromo plot Outputs

Use chromo plot outputs to visually check MUMmer coords or minimap2 PAF alignments without re-running the aligner. For examples of how to read clean, reversed, chimeric, inverted, duplicated, gapped, and noisy patterns, see How to Interpret Dot Plots. When --sel-ref is used, the whole-genome output is restricted to the selected references and --per-ref writes only those selected reference panels.

Table 7. Example plot file set. A single run can write a whole-genome plot and, with --per-ref, one plot per reference sequence in each requested format.

chromo graph-map Outputs

Use chromo graph-map outputs when a unitig-level GFA must be projected onto contig FASTA coordinates before graph evidence can be compared to dot plots or breakpoint candidates.

chromo scaffold Outputs

Use chromo scaffold outputs after final sorting and review to inspect scaffold FASTA construction, inferred gaps, overlap handling, and optional graph junction evidence.

Table 8. Example scaffold_gaps.tsv row. The gap report records the flanking contigs, inferred reference-space gap, FASTA gap actually written, overlap classification, and overlap policy action.

Table 9. Example scaffold_summary.tsv rows. The summary table gives one row per emitted scaffold record.

Listing 5. Example scaffold FASTA output. Scaffold headers summarize the number of source contigs, sequence bases, gap bases, and gap mode.

>chr1 contigs=2 sequence_bp=7 gap_bp=5 gap_mode=inferred
AAAANNNNNTTT
>chr2 contigs=1 sequence_bp=2 gap_bp=0 gap_mode=inferred
GG

chromo gapfill Outputs

Use chromo gapfill outputs to review graph-supported candidate fills before optional sequence-changing application.

Patch comparison columns are populated when --patch-table is provided: patch_candidate_count, patch_best_id, patch_best_source, patch_best_bp, patch_graph_status, patch_best_notes, and, with --include-fill-sequences, patch_best_sequence. patch_graph_status values include exact_graph_match, same_length_graph_mismatch, graph_mismatch, and patch_only_no_graph_sequence.

Table 10. Example gapfill_plan.tsv row. Selected columns from a graph fixture show a junction resolved by reference-placement PAF support. The default accept_fill=no makes planning review explicit before strict reviewed application.

Listing 6. Example applied gapfilled FASTA output. With --apply --reviewed-plan, accepted fillable paths insert graph sequence and trim the right flank by the terminal GFA overlap; unresolved or unaccepted junctions use fallback N gaps.

>chr1 contigs=2 filled_gaps=1 fallback_gaps=0 fill_bp=4 fallback_gap_bp=0 trimmed_bp=4
AAAACCCCGGGGTTTT

Audit Tables

The report tables are deliberately redundant. They keep original contig IDs, new contig IDs, reference assignments, coordinates, status labels, and decision metrics close to the sequence output they explain. Prefer using these tables instead of parsing FASTA headers when downstream scripts need assignment, split, cut, scaffold, or gap metadata.

Sequence-Changing Outputs

Only a subset of commands change sequence:

• chromo fix replaces selected contigs with split pieces.
• chromo cut replaces selected contigs with pieces cut at exact reviewed positions.
• chromo manual apply reproduces a reviewed browser recipe.
• chromo scaffold joins sorted contigs with inferred or fixed N gaps and only trims overlaps when an explicit overlap policy asks it to.
• chromo gapfill --apply --reviewed-plan inserts graph sequence only for fillable and accepted paths; --apply --apply-all-fillable is the explicit exploratory all-fillable mode.

chromo eval, chromo manual, chromo plot, and report-only graph or long-read evidence modes do not trim, polish, or rewrite sequence. chromo sort does rewrite retained records into an ordered FASTA and can optionally orient them with --orient-to-reference.