Studying variation at specific loci across populations or species usually
means either building an expensive whole-genome graph or falling back to
a single reference. impg takes a third path: it treats all-vs-all
pairwise alignments as an implicit pangenome graph and projects
target ranges through the alignment network to extract only the
homologous sequences you need. Query regions across hundreds of
genomes in seconds, walk transitive alignments, partition a cohort into
comparable loci, refine regions to maximize sample support — all
without ever materializing a graph structure.
At its core, impg lifts ranges from a target sequence (the reference
in a given alignment) into the queries aligned onto it. It outputs
BED / BEDPE / PAF — ready to feed FASTA extraction, multiple sequence
alignment, or a graph builder like pggb or minigraph-cactus — and
can also emit GFA directly by chaining sweepga + seqwish + smoothxg-style
smoothing.
impg also embeds syng as a second backend: build a syncmer GBWT index once from a FASTA or AGC, then query / partition against it without ever running an aligner. Shared syncmers act as anchors; transitive chains define homology. Same output formats as the alignment path. impg map projects short reads onto a syng index in GAF or PAF form.
impg uses coitrees (cache-oblivious
interval trees) for fast range lookup, and stores CIGAR strings as
compact deltas. The result is fast, memory-efficient projection of
sequence ranges through alignment networks.
# Bioconda
conda install -c bioconda impg
# Source
git clone --recursive https://github.com/pangenome/impg.git
cd impg
cargo install --force --path .The source install places impg, gfaffix, and the companion aligner
binaries (wfmash, FastGA) into ~/.cargo/bin/.
docker pull pangenome/impg
docker run pangenome/impgOn older glibc systems (e.g. Debian Buster) a plain cargo build can
fail because wfmash needs modern CMake / GCC / glibc. Use Guix's
toolchain:
source ./env.sh
cargo build --releaseSee .guix/ for the Guix build recipe (guix build -L .guix/modules --file=guix.scm). For libclang link errors, set LIBCLANG_PATH to
your LLVM install (see env -i … LIBCLANG_PATH=…).
impg query -a cerevisiae.pan.paf.gz -r S288C#1#chrI:50000-100000 -d 1000 -x-a— alignment file (PAF / 1ALN / TPA). PAF must use=/XCIGAR ops (fromwfmashorminimap2 --eqx).-r— target range,seq:start-end.-d— merge query-gathered ranges separated by up to this many bp. This is also the largest internal gap/SV one query hop can absorb into one reported interval.-x— walk the transitive closure: find everything aligned to the initial result, recursively.
Example output (BED):
S288C#1#chrI 50000 100000
DBVPG6044#1#chrI 35335 85288
Y12#1#chrI 36263 86288
DBVPG6765#1#chrI 36166 86150
YPS128#1#chrI 47080 97062
UWOPS034614#1#chrI 36826 86817
SK1#1#chrI 52740 102721
All commands accept -a (alignment files, mixed PAF/1ALN/TPA) or
--alignment-list (text file, one per line), -t / --threads, and
-v 0|1|2 for verbosity. Every command has a --help with the
exhaustive flag list — this section covers the flags you'll actually turn.
# A single range
impg query -a aln.paf -r chr1:1000-2000 -d 100
# Transitive closure (depth 2 by default)
impg query -a aln.paf -r chr1:1000-2000 -d 100 -x -m 3
# Many regions from a BED, mixed PAF + 1ALN
impg query -a f1.paf f2.1aln -b regions.bed -d 100
# One local graph per BED4 row, named from column 4
impg query -a pan.syng -b regions.bed -d 100k -o gfa:syng:crush \
--sequence-files genomes.agc -O graphs/
# Output formats: auto | bed | bedpe | paf | gfa | vcf | maf | fasta | fasta+paf | fasta-aln
impg query -a aln.paf -r chr1:1000-2000 -d 100 -o bed
impg query -a aln.paf -r chr1:1000-2000 -d 100 -o gfa --sequence-files genomes.fa
impg query -a aln.paf -r chr1:1000-2000 -d 100 -o vcf --sequence-files genomes.fa
impg query -a aln.1aln -r chr1:1000-2000 -d 100 -o fasta --sequence-files *.fa \
--reverse-complement
# Filter / shape the result
impg query -a aln.paf -r chr1:1000-2000 --min-identity 0.9 -l 5000 -d 1000
# Restrict to a sequence whitelist (also filters transitive intermediates)
impg query -a aln.paf -r chr1:1000-2000 -d 100 -x --subset-sequence-list seqs.txt
# Fast approximate mode (.1aln only; bed/bedpe output)
impg query -a aln.1aln -r chr1:1000-2000 -d 100 --approximate
# Alignment-free path: -a accepts a syng index prefix (see `syng` below)
impg query -a pan.syng -r chr1:1000-2000 -d 100 --sequence-files genomes.faGFA / MAF / FASTA outputs need --sequence-files (FASTA or AGC
archive) or --sequence-list. See GFA engines for
engine selection and partitioned builds. With -b and graph-like outputs
(gfa, vcf, or gbwt), -O is treated as an output directory and
each BED row is written separately using the BED column 4 name as the
file stem.
Runs alignment + seqwish + (optional) smoothing, no pre-computed alignment needed.
# Default pipeline: pggb (align → seqwish → smooth → gfaffix)
impg graph --sequence-files genomes.fa -g output.gfa -t 16
# Partitioned mode for large inputs (aligns once, then builds per-window)
impg graph --sequence-files genomes.fa -g output.gfa --gfa-engine pggb:10000
# Reuse an existing PAF instead of aligning
impg graph --sequence-files genomes.fa -g output.gfa --paf-file aln.paf
# Batch alignment to cap per-batch RAM (wfmash) or disk (FastGA)
impg graph --sequence-files genomes.fa -g output.gfa --batch-bytes 2Gquery -o gfa and graph share the same engine code and flags — the
only difference is where the sequences come from (IMPG index +
sequence files for query; FASTAs directly for graph).
# 1Mb windows, single BED output
impg partition -a aln.paf -w 1000000 -d 100000
# One FASTA per partition (for downstream pipelines)
impg partition -a aln.1aln -w 1000000 -d 100000 -o fasta --sequence-files *.fa \
--separate-files --output-folder partitions/
# Selection strategies pick the next starting sequence
impg partition -a aln.paf -w 1000000 -d 100000 --selection-mode longest # default
impg partition -a aln.paf -w 1000000 -d 100000 --selection-mode sample # PanSN sample
impg partition -a aln.paf -w 1000000 -d 100000 --selection-mode haplotype # PanSN haplotype
# Start from a fixed list of sequences
impg partition -a aln.paf -w 1000000 -d 100000 --starting-sequences-file seqs.txt
# GFA output per partition; engines: pggb | seqwish | poa
impg partition -a aln.paf -w 1000000 -d 100000 -o gfa --gfa-engine pggb \
--sequence-files *.fa --separate-files --output-folder gfas/
# Fully partitioned pipeline: build → lace → one gfaffix pass
impg partition -a aln.paf -w 100000 -d 100000 -o gfa --gfa-engine pggb:10000 \
--sequence-files *.fa --output-folder results/Explores asymmetric left/right expansions around each range, picking the smallest window that keeps the most sequences, samples, or haplotypes fully spanning it. Useful for anchoring loci outside structural variants.
impg refine -a aln.paf -r chr1:1000-2000
impg refine -a aln.paf -b loci.bed --span-bp 2000 -d 200000
# Maximize PanSN samples / haplotypes instead of raw sequence count
impg refine -a aln.paf -r chr1:1000-2000 --pansn-mode sample
impg refine -a aln.paf -r chr1:1000-2000 --pansn-mode haplotype
# Cap expansion distance
impg refine -a aln.paf -r chr1:1000-2000 --max-extension 0.90 # 90% of locus
impg refine -a aln.paf -r chr1:1000-2000 --max-extension 50000 # 50kb absolute
# Emit the spanning-entity list alongside the refined BED
impg refine -a aln.paf -r chr1:1000-2000 --support-output support.bedimpg similarity -a aln.paf -r chr1:1000-2000 --sequence-files *.fa
impg similarity -a aln.1aln -b regions.bed --sequence-files *.fa --distances
# Group by PanSN prefix
impg similarity -a aln.paf -r chr1:1000-2000 --sequence-files *.fa \
--delim '#' --delim-pos 2 # sample#haplotype
# PCA / MDS on the distance matrix
impg similarity -a aln.paf -r chr1:1000-2000 --sequence-files *.fa \
--pca --pca-components 3 --pca-measure cosine# GFAs (auto-detected)
impg lace -f gfa1.gfa gfa2.gfa gfa3.gfa -o combined.gfa
# From a list file, fill inter-window gaps with sequence
impg lace -l gfa_list.txt -o combined.gfa --fill-gaps 1 --sequence-files ref.fa
# VCFs
impg lace -f *.vcf -o combined.vcf --reference ref.faPath names must follow NAME:START-END (e.g.
HG002#1#chr20:1000-2000); the coordinates drive reassembly. NAME
may contain : — the last : is the separator.
Recommended post-processing:
gfaffix combined.gfa -o combined.fix.gfa &> /dev/null
odgi unchop -i combined.fix.gfa -o - -t 16 | \
odgi sort -i - -o - -p gYs -t 16 | \
odgi view -i - -g > combined.final.gfa# Single combined index
impg index -a aln.paf -i aln.impg
# Mixed PAF + 1ALN + TPA
impg index -a f1.paf f2.1aln f3.tpa -i all.impg
# Per-file index (faster incremental rebuilds for large cohorts)
impg index --alignment-list files.txt --index-mode per-file--index-mode auto (default) picks per-file when ≥ 100 files are
listed, single otherwise. impg warns when the index is older than its
input alignments; -f/--force-reindex rebuilds.
Bgzipped PAFs work natively (impg reads .paf.gz); optionally
bgzip -r alignments.paf.gz creates a .gzi sidecar to speed up the
first read.
impg stats -a aln.paf
impg stats -a f1.paf f2.1alnimpg syng builds a syng index from FASTA or AGC. Six sidecars are written under one prefix (<prefix>.1khash dictionary, .1gbwt GBWT, .syng.names, .syng.pstep sampled path-step checkpoints, .syng.spos sampled syncmer occurrence positions, .syng.meta parameters — auto-loaded on read). Any later impg query / partition / map can then point -a at the prefix (or any sidecar) and skip pairwise alignment. Exact target coordinates are located by walking GBWT occurrences forward to the next sampled checkpoint and resolving it through .syng.spos.
Parameters follow the syng paper: --smer-length (s, default 8) and --syncmer-length (k, must be odd, default 63). Position sidecars use a regular per-path syncmer-step grid plus the terminal syncmer: --position-sample-rate 256 samples steps 0, 256, 512, ... and the final step on each path. --parallel-dictionary adds a deterministic prepass for large inputs.
impg map projects FASTA/FASTQ queries onto a syng index via shared syncmers. The default output is GAF (per-read syncmer-node walks); pass -o paf for projected genome coordinates, -o pack for a compact binary node support vector, -o pack-tsv for a human-readable TSV support vector, or -o proj for a sample projection bundle containing both sample.pack and reads.gaf.zst. Text map output written with -O is compressed automatically when the filename ends in .zst or .zstd; pack uses internal block zstd compression for random access by node ID. packbin remains accepted as a compatibility alias for compact pack output.
Use impg syng-repair -a <prefix> --position-sample-rate <N> --force to rebuild or resample .syng.pstep and .syng.spos from an existing .1gbwt / .1khash syng index without re-reading the original sequences.
End-to-end walkthrough using ODGI's C4 test GFA (90 HPRC haplotypes, ~6.9 Mb total):
# 1. Get the C4 GFA from odgi/test/, dump paths to FASTA
curl -sL -O https://raw.githubusercontent.com/pangenome/odgi/master/test/chr6.C4.gfa
odgi paths -i chr6.C4.gfa -f > chr6.C4.fa
samtools faidx chr6.C4.fa
# 2. Build the syng index (~80 ms on this dataset, 4 threads)
impg syng -f chr6.C4.fa -o c4.syng \
--syncmer-length 63 --smer-length 8 --syncmer-seed 7 \
--position-sample-rate 256 -t 4
# writes c4.syng.1khash / .1gbwt / .syng.names / .syng.spos / .syng.pstep / .syng.meta
# 3. Query a 10 kb grch38 sub-window onto all 90 haplotypes
impg query -a c4.syng -r 'grch38#chr6:31972046-32055647:0-10000' \
-d 150 --sequence-files chr6.C4.fa | head -3
# HG00673#1#JAHBBZ010000030.1:31835924-31919525 0 10000 grch38...:0-10000 . +
# HG01123#2#JAGYYY010000050.1:31954985-32038586 0 10001 grch38...:0-10000 . +
# HG01243#1#JAHEOY010000117.1:3252171-3329407 0 9999 grch38...:0-10000 . +
# 4. Whole grch38 C4 path (length from .fai) → 36 hits, FASTA out
LEN=$(awk -F'\t' '$1=="grch38#chr6:31972046-32055647"{print $2}' chr6.C4.fa.fai)
impg query -a c4.syng -r "grch38#chr6:31972046-32055647:0-${LEN}" \
-d 150 --sequence-files chr6.C4.fa -o fasta > c4.homologs.fa
# 5. Map a 2 kb probe back onto the index
samtools faidx chr6.C4.fa 'grch38#chr6:31972046-32055647:5000-7000' > probe.fa
impg map -a c4.syng -q probe.fa | cut -f1-6 | head -1
# grch38...:5000-7000 2001 16 1979 + <264>265>266<267<268>...
impg map -a c4.syng -q probe.fa -o pack-tsv | head
# #node_id count
# 264 1
# 265 1
impg map -a c4.syng -q probe.fa -o pack-tsv -O probe.pack.tsv.zst
impg map -a c4.syng -q probe.fa -o pack -O probe.pack \
--pack-compression-level 12
impg map -a c4.syng -q probe.fa -o proj -O probe.proj
impg map -a c4.syng -q probe.fa -o paf
# grch38...:5000-7000 2001 1302 1833 + HG02109#1#... 77232 6300 6831 126 531 0 an:i:2 sk:i:63pack stores a dense u8 count vector over syng node IDs, zstd-compressed
in independently addressable blocks, plus a small overflow table for node
counts above 255. --pack-compression-level defaults to 12; level 19 is a
compact cohort/archive setting that is slower to write but just as fast to
read in practice.
A proj directory is the richer sample projection object:
sample.proj/
manifest.json
sample.pack
reads.gaf.zst
Use pack when aggregate graph support is enough or when storing many
samples. Use proj when inference should also have read-walk linkage evidence
available.
impg genotype (impg gt) is the namespace for graph-based genotyping
methods. The first listed method is cos: cosine genotyping over graph-feature
coverage, following the COSIGT/LikeGT scoring model:
impg genotype cos -a c4.syng -p sample.pack \
-r 'grch38#chr6:31972046-32055647:0-10000' \
--ploidy 2 --top-n 20cosigt is accepted as an alias for cos to make the paper/tool lineage
clear. cos extracts haplotype candidates from the requested reference path
range, builds traversal-count vectors over the candidate syncmer nodes, and ranks
ploidy-sized haplotype combinations against the sample pack
coverage vector by cosine similarity. The default candidate mode is
spanning: candidates must have shared anchors spanning the requested
reference interval. Use --candidate-mode overlapping to score each gathered
candidate interval independently. See docs/genotype-architecture.md for the
graph-feature evidence model this is built around.
impg infer lifts the same pack/cos scoring path over ranges or partitions and
emits allele calls over genomic intervals:
impg infer -a hprc.syng --proj sample.proj \
--partitions partitions.bed --top-n 1 -O sample.infer.tsv.zstInputs choose the operation mode: -r/--target-range types one range,
--target-bed types a BED-like range list, --partitions consumes ready
impg partition BED output, and omitting all three runs internal syng partition
discovery, which requires -d/--merge-distance. infer can consume --pack
for aggregate graph support or --proj for a bundle that also carries read
walks. Add --stitch beam --emit-mosaic out.tsv --emit-fasta haps.fa --emit-gfa diplotype.gfa --sequence-files panel.fa-or.agc to stitch local
calls into phased mosaic paths and materialize sequence. When --proj or
--gaf read walks are available, stitched inference rewards phase transitions
supported by the same reads (--read-link-weight,
--min-read-link-anchors) while --switch-penalty keeps unnecessary panel
crossovers expensive. Add --phase-block-size N to split broad targets into
internal phase blocks before stitching, which lets the same model infer copied
segments inside a partition. See docs/infer-design.md.
-r splits on the last :. Path names from odgi paths -f
already contain coordinates (grch38#chr6:31972046-32055647), so a
whole-sequence query still needs an explicit trailing sub-range
(:0-${LEN}) — otherwise impg parses grch38#chr6 as the sequence
and the bare 31972046-32055647 as the range, and the genome lookup
fails.
Default query mode runs BiWFA boundary refinement and so requires
--sequence-files. Pass --syng-raw for the raw syncmer-resolution
pass-through. Tune via --syng-padding, --syng-min-chain-anchors
(lower -> more paralog hits; default is an adaptive cap derived from
query length, syncmer density, and expected exact-syncmer survival at
95% identity), --syng-min-chain-fraction (default 0.5; set 0 for
exploratory local-chain discovery), and the seed-frequency filters
--syng-seed-max-occurrences / --syng-seed-drop-top-fraction.
By default syng query drops the top 0.05% most frequent query-local seed
syncmers, seeds ranges from bounded exact GBWT walk seeds of five
syncmers (--syng-seed-walk-anchors; set 3 for more sensitive seeds),
and does not use a fixed absolute occurrence cap; set
--syng-seed-max-occurrences only when a hard panel-specific ceiling is
desired. High-frequency syncmers are ignored only while seeding candidate
ranges; once a range survives, downstream scoring can still recover all
syncmers by walking that path range. query -o gbwt emits a
region-specific sub-GBWT. The syng prefix passed to -a is resolved
relative to cwd, so either cd to the index directory or pass an absolute
path. For a focused syng-backed local GFA recipe, see
docs/syng-gfa-query.md.
For graph outputs, query -o gfa|vcf --render-graph also renders the final
1D graph with gfalook. The default image is PNG beside the -O output prefix
(<prefix>.png); --render-graph-output overrides the image path, and
--render-graph-format svg or a .svg suffix emits SVG. Add
--render-graph-depth to pass -m to gfalook for mean-depth coloring. With
-b regions.bed graph output, -O is a directory for the graph files and the
render output is written per BED row using the sanitized BED column 4 name.
If you already have a local GFA, call variants directly with POVU via
impg gfa2vcf -g local.gfa -o local.vcf -r ref_path. This is the same
GFA-to-VCF conversion used internally by query -o vcf and partition -o vcf.
The graph, query -o gfa, and partition -o gfa commands share
one set of engine implementations, selected via --gfa-engine:
| Engine | Pipeline | Use for |
|---|---|---|
pggb (default) |
sweepga + seqwish + smoothxg-style smoothing + gfaffix | smoothed variation graphs |
seqwish |
sweepga + seqwish + gfaffix | raw (unsmoothed) graphs |
poa |
single-pass SPOA | small regions, quick MSA-based output |
syng / syng:blunt |
regional syng syncmer graph with exact zero-overlap path materialization | source-spelling syng graph output from syng indexes |
syng:raw |
regional syng syncmer overlap graph | explicit native overlap graph for debugging or overlap-aware consumers |
syng-local / syng-local:blunt |
extract query-selected sequences, build a fresh local syng graph, then exact zero-overlap materialization | experimental regional syncmer parameter sweeps |
For syng-index queries, --gfa-engine syng defaults to syng:blunt.
The compact form -o gfa:syng:blunt,k=63,s=8,seed=7 is accepted as
shorthand for -o gfa --gfa-engine syng:blunt,k=63,s=8,seed=7; the
k/s/seed tail is checked against the loaded syng index.
Blunt syng output is sequence-preserving when the graph can fetch source DNA
for non-syncmer spans via --sequence-files or the local temporary FASTA used
by syng-local. Without source sequence files, missing gap DNA is emitted as
N, and those paths are explicitly not source-preserving. Use syng:raw when
you want the native syng overlap graph; raw path spelling requires an
overlap-aware parser and is not the same as concatenating S-line DNA.
Use syng-local when the local graph should be rebuilt from the extracted
regional sequences with its own syncmer scheme, for example
-o gfa:syng-local:blunt,k=127,s=16,seed=7:crush; in this mode k/s/seed
select the local rebuild parameters rather than asserting the global index.
Unlike syng, syng-local does not apply the syng frequency mask unless
:mask is requested explicitly.
Syng GFA extraction selects the top 0.05% most frequent local syncmer nodes and
private-splits unsupported high-frequency occurrences before raw or blunt graph
materialization. Occurrences in supported high-frequency runs or exact spans
stay shared (freq-run=10 and freq-span=1k by default). The same
run/span rescue is applied to spectrum-selected dispersed scaffold-glue nodes,
so there is no second private-split path that bypasses the high-frequency
support policy. It also splits rare repeated-copy local syncmer contexts by
default, so high-copy repeats and single-syncmer repeat loops do not become
global graph glue. Use
-o gfa:syng:nomask to disable this, or
-o gfa:syng:mask,top=0.001,max-occ=500,freq-run=10,freq-span=1k:crush to
tune it. Use freq-run-aware=false or legacy-freq-mask=true to reproduce the
older node-level frequency removal for debugging.
Add :cut-ns, for example -o gfa:syng:cut-ns:crush, to drop assembly N-runs
from fetched gap DNA and split graph paths at those breaks.
For graph builds from query-extracted sequences, add an explicit terminal
N-run clipping stage before the engine: -o gfa:cut-n=100:pggb. This clips
only leading and trailing N/n runs whose length is at least the requested
threshold before graph construction; internal N-runs and shorter terminal
N-runs are preserved. There is no default clipping threshold, and omitting
cut-n=<bp> leaves extracted sequences unchanged.
The same shorthand works for the other engines: -o gfa:pggb,
-o gfa:seqwish, -o gfa:poa, and -o gfa:syng. Alignment-backed graph
builds may also include the aligner prefix, for example
-o gfa:wfmash:seqwish, -o gfa:fastga:pggb, or
-o gfa:sweepga:seqwish; this is equivalent to setting --aligner and
--gfa-engine separately.
For a C4/HPRCv2-style local render where terminal assembly N blocks would otherwise become noisy graph tips:
impg query -a ~/hprcv2/HPRC_r2_assemblies_0.6.1.syng \
-r GRCh38#0#chr6:<C4-range> \
--sequence-files ~/hprcv2/HPRC_r2_assemblies_0.6.1.agc \
-d 100000 \
-o gfa:cut-n=100:pggb \
-O c4.cutn100.pggb.gfaVCF output uses the same graph engines and then converts the resulting local
GFA through POVU. Use -o vcf --gfa-engine <engine> or the shorthand
-o vcf:<engine>, for example -o vcf:syng.
Append :WINDOW to any engine to build per-window and lace:
impg query -a aln.paf -r chr1:0-500000 -o gfa \
-d 1000 --gfa-engine pggb:10000 --sequence-files *.fa -O out
impg graph --sequence-files *.fa -g out.gfa --gfa-engine seqwish:10000
impg partition -a aln.paf -w 100000 -d 100000 -o gfa --gfa-engine pggb:10000 \
--sequence-files *.fa --output-folder results/Window size is in bp (≥ 1000). Partitioned mode is the recommended approach for large regions — it caps peak memory and runs one final gfaffix pass over the laced graph.
The flags below are available on all three GFA-producing commands. Defaults match pggb's conventions; only tune if the default graph doesn't meet your need.
# Seqwish induction
--min-match-len 23 # minimum transitive-match length
--transclose-batch 10000000 # batch size (reduce for lower memory)
--sparse-factor 0.0 # drop this fraction of input matches
--disk-backed # use disk-backed interval trees
--repeat-max / --min-repeat-dist
# Smoothxg-style smoothing (pggb only)
--target-poa-length 700,1100 # one pass per value
--max-node-length 100
--poa-padding-fraction 0.001
# Alignment filtering (sweepga, seqwish + pggb only)
--no-filter # skip post-alignment filtering
--num-mappings many:many # plane-sweep cardinality
--scaffold-jump 50000 # scaffold chaining gap (0 = off)
--scaffold-mass 10000 # min scaffold chain length
--overlap 0.95
--min-aln-identity 0.9
# Aligner backend
--aligner wfmash # default; alt: fastga
--sparsify auto # wfmash-only; pair-selection heuristic
--map-pct-identity 90 # wfmash -p value
--fastga-frequency / --fastga-frequency-multiplier # fastga-only
# Temp files (can be large)
--temp-dir /scratch/tmp # explicit path
--temp-dir ramdisk # → /dev/shm on LinuxCombining --aligner fastga with --sparsify or --aligner wfmash
with --fastga-frequency is rejected at parse time.
-a / --alignment-files— one or more PAF/1ALN/TPA files (can be.gz).--alignment-list— text file, one alignment path per line.-i / --index— existing IMPG index.-f / --force-reindex— rebuild even if the index is up-to-date.-t / --threads— default4.-d / --merge-distance— required for query/partition/refine/similarity: merge query-gathered ranges within this gap (bp). This is the largest internal gap/SV one query hop can absorb into one reported interval.--no-merge— disable merging.--consider-strandness— keep strands separate during merge.--subset-sequence-list— restrict results to listed sequences.--unidirectional— disable bidirectional alignment interpretation.
Sequence-requiring outputs (GFA/MAF/FASTA, similarity, lace --fill-gaps) take --sequence-files (FASTA or AGC) or
--sequence-list.
FASTA="cerevisiae.fa.gz"
PAF="cerevisiae.paf"
THREADS=16
# 1. Index
impg index -a "$PAF" -i yeast.impg -t "$THREADS"
# 2. Partition into 100kb windows, one FASTA per window
mkdir -p partitions gfas
impg partition -i yeast.impg -w 100000 \
-d 100000 --sequence-files "$FASTA" -o fasta \
--separate-files --output-folder partitions -t "$THREADS"
# 3. Build per-partition GFAs in parallel
ls partitions/*.fasta | xargs -P 4 -I {} bash -c '
f="{}"; base=$(basename "$f" .fasta)
impg graph --sequence-files "$f" -g "gfas/${base}.gfa" -t 4
'
# 4. Lace, filling inter-window gaps with reference sequence
find gfas -name "*.gfa" -size +0 | sort -V > gfa_list.txt
impg lace --file-list gfa_list.txt --sequence-files "$FASTA" \
-o yeast.gfa --fill-gaps 2 -t "$THREADS"
# 5. Post-process with odgi
odgi build -g yeast.gfa -o yeast.og -t "$THREADS"
odgi sort -i yeast.og -o yeast.sort.og -O -p Ygs -t "$THREADS"
odgi layout -i yeast.sort.og -o yeast.lay -t "$THREADS"
odgi viz -i yeast.sort.og -o yeast.viz.png -x 4000 -y 1000 -s '#'
odgi draw -i yeast.sort.og -c yeast.lay -p yeast.draw.pngFor modern inputs, you can replace steps 1–4 with a single
impg graph --sequence-files "$FASTA" -g yeast.gfa --gfa-engine pggb:100000.
scripts/faln2html.py renders the fasta-aln output into an
interactive HTML MSA using react-msa
or ProSeqViewer.
impg query -a aln.paf -r chr1:1000-2000 -d 100 -o fasta-aln --sequence-files *.fa \
| python scripts/faln2html.py -i - -o alignment.html [--tool proseqviewer]Andrea Guarracino aguarra1@uthsc.edu · Bryce Kille brycekille@gmail.com · Erik Garrison erik.garrison@gmail.com
MIT.