Open-source desktop tool for viewing, editing and analyzing Sanger sequencing
chromatograms (.ab1 files) — a free, offline alternative to FinchTV / Chromas
for everyday molecular biology lab work. English/Russian interface.
Viewer
- Parse ABIF (
.ab1): traces, phred quality, peak positions, base calls - Smooth chromatogram viewer: zoom/scroll (wheel, keyboard, Ctrl+drag to zoom into a selection), position ruler, quality bars, per-base tooltips
- Whole-read minimap with quality map and optional quality profile
- Dark theme, per-channel (A/C/G/T) visibility, trace smoothing
- Find subsequence (Ctrl+F, IUPAC-aware, both strands), go to base (Ctrl+G)
Editing & trimming
- Manual base editing (change / insert / delete) with unlimited undo/redo
- Reverse-complement as a single undoable operation
- Quality trimming (sliding window, Chromas-style) with live preview, draggable trim handles
- FASTA / FASTQ (phred33) / PNG / PDF export; editing sessions are saved and restored automatically
Batch mode
- Scan a folder of
.ab1files: per-file QC (length, mean Q, Q20, trim length), status marks, sorting, hiding bad reads - Batch trim + export respecting per-file editing sessions; CSV summary
Analysis
- Reference alignment: exact semiglobal DP with affine gaps (Bio.Align.PairwiseAligner, free end gaps — read ends are never dropped), both read orientations, ambiguity flag for repetitive references
- SNP calling: homozygous vs heterozygous (double-peak detection by secondary/primary height ratio, IUPAC codes), low-quality flagging
- Expected-SNP verification: load a VCF/CSV/TSV of expected variants and get a verdict per position (confirmed / heterozygous / not found / no data / low quality / input-error guards), with marks on the chromatogram and CSV / Excel reports
- Send the current read to NCBI BLAST (opens results in the browser)
Check a set of plasmid or amplicon clones against the expected construct from
their Sanger reads: open the clone-verification workspace, load the reference (FASTA,
GenBank or SnapGene), drop the .ab1 reads of the whole plate, optionally set
the region of interest, and read the verdict per clone off the colour matrix.
Scenario. Load the construct, add the reads (files or a folder), check the
automatic grouping into clones, choose the region; every clone gets its verdict right away. Click a
difference in the matrix to see the stacked chromatograms of every read at that
position. Export the result as Excel/CSV, or save the whole session as an
.ab1proj file. FASTA constructs are treated as circular (reads may span the
origin); GenBank/SnapGene files take the topology from the file; the Circular
checkbox (--linear on the command line) overrides the topology.
File names. Reads are grouped by name as <clone>_<primer>: the primer is
the last known primer token (F, R, T7, SP6, M13F, pUC-F, ...); if
the name has none, the last token is taken as the primer. The clone is what
precedes the primer, minus a single-digit order token and a leading plate well
(A05), the well being stripped only when another clone token follows it
(B2_F is clone B2). A05_clone3_2_R.ab1 is clone clone3, primer R;
Achl_ACHLO006-09_1_F.ab1 is clone Achl_ACHLO006-09. Files whose name yields
no clone stay unassigned and are skipped. For other conventions give a regular
expression with named groups clone and primer that matches the whole file
name without the extension, e.g. (?P<clone>\d+)-(?P<primer>[A-Za-z0-9]+).
Verdicts. Checked in this order, the first match wins:
| Verdict | Meaning |
|---|---|
no_data (grey) |
no read of the clone aligned to the construct |
mixed (purple) |
at least 3 positions (default) with a double peak: probably a mixed colony |
errors (red) |
at least one difference in the region that is confirmed or disputed |
incomplete (yellow) |
no such difference, but part of the region is not covered by any read |
ok (green) |
the region is fully covered and shows no reliable difference |
A difference is confirmed (only supporting reads, Q >= min Q), disputed
(another read shows the reference there) or low quality (seen only below
min Q). Both confirmed and disputed differences give errors; only low-quality
ones do not change the verdict. Defaults: min Q 20, double-peak ratio 0.40.
Command line (no GUI, no Qt needed):
ab1kit verify-clones --construct pET28-gene.gb --region 120-1850 reads/ -o report.xlsx
ab1kit qc reads/ -o qc.csv
ab1kit export reads/ --trim --fasta -o trimmed/
verify-clones prints a table of clones and exits with 0 if at least one clone
is ok, 1 if none is, 2 on invalid input; -o report.csv writes report.csv
and report_differences.csv. The differences count (table and report) is the
in-region differences that are not low quality, i.e. those behind errors.
Reads that could not be used (unreadable, not aligned, ambiguous placement) are
listed on stderr as ab1kit: <file>: <reason> and in the read_issues column. Options: --region A-B (1-based, inclusive),
--linear, --pattern REGEX, --min-q N, --het-ratio X.
The exit-code contract applies to the ab1kit command installed with pip. The
Windows release ab1kit.exe is a windowed program: the shell returns to the
prompt immediately, so exit codes and output ordering are unreliable and output
may appear after the prompt. Wait for it explicitly: start /wait ab1kit.exe verify-clones ... (cmd), or Start-Process -Wait -NoNewWindow /
(Start-Process ab1kit.exe -ArgumentList ... -Wait -PassThru).ExitCode
(PowerShell).
Requires Python 3.11+:
python -m venv .venv
.venv/Scripts/pip install -e ".[dev]" # Windows; on POSIX: .venv/bin/pip
.venv/Scripts/python -m pytest # run the test suite
.venv/Scripts/python -m pytest -m perf # timing tests (skipped by default)
.venv/Scripts/ab1kit # launch the GUI
ab1kit also accepts a file or folder argument. Prebuilt binaries are on the
Releases page:
- Windows:
ab1kit.exe(single file, no installation). - Linux (x86_64):
ab1kit-linux-x86_64.tar.gz. Unpack withtar -xzf ab1kit-linux-x86_64.tar.gzand run./ab1kit. Built on Ubuntu 22.04, so it needs glibc 2.35 or newer. The GUI needs the usual Qt runtime libraries; on Debian/Ubuntu:sudo apt install libegl1 libxkbcommon-x11-0 libxcb-cursor0 libxcb-icccm4 libxcb-image0 libxcb-keysyms1 libxcb-randr0 libxcb-render-util0 libxcb-shape0 libxcb-xinerama0 libxcb-xfixes0 libgl1 libfontconfig1 libdbus-1-3 libglib2.0-0 libxcb-xkb1 libxcb-xinput0 libxcb-shm0. The Linux binary prints CLI output to the terminal normally.
| Shortcut | Action |
|---|---|
| Ctrl+O / Ctrl+Shift+O | open file / folder |
| wheel, Ctrl+drag | zoom / zoom to selection |
| ←→ PgUp/PgDn Home/End | navigate |
| + / − | zoom in / out |
| Ctrl+G | go to base |
| Ctrl+F / F3 | find subsequence / next match |
| Ctrl+Z / Ctrl+Y | undo / redo |
| Ctrl+R | reverse-complement |
| Ctrl+T | quality trimming |
| Ctrl+S | save editing session |
| Ctrl+L | load reference |
| Ctrl+Shift+V | check expected SNPs |
Python 3.11+, PySide6 (Qt) GUI, BioPython for ABIF parsing and alignment.
Core logic (chrom_core) is UI-independent and covered by pytest. MIT
license — see LICENSE.
Issues and pull requests are welcome at
github.com/adrobi/ab1kit.
Example chromatograms in examples/ come from the
BioPython test suite
(Biopython License).
v1.1.0 — adds clone verification, a command-line interface, a refreshed UI and more accurate peak analysis on top of the 1.0 viewer, editing, batch mode, alignment, SNP calling and expected-SNP verification. See CHANGELOG.md.