BioinformaticsBioinformaticsPeer-reviewed journal article
clinker & clustermap.js: Automatic generation of gene cluster comparison figures
Abstract
SummaryGenes involved in biological pathways are often collocalised in gene clusters, the comparison of which can give valuable insights into their function and evolutionary history. However, comparison and visualisation of gene cluster homology is a tedious process, particularly when many clusters are being compared. Here, we present clinker, a Python based tool, and clustermap.js, a companion JavaScript visualisation library, which used together can automatically generate accurate, interactive, publication-quality gene cluster comparison figures directly from sequence files. Availability and ImplementationSource code and documentation for clinker and clustermap.js is available on GitHub (github.com/gamcil/clinker and github.com/gamcil/clustermap.js, respectively) under the MIT license. clinker can be installed directly from the Python Package Index via pip. ContactE-mail: cameron.gilchrist@research.uwa.edu.au, yitheng.chooi@uwa.edu.au
- assay: protein sequence
- research focus: software / tool
- research focus: phylogenetics / evolution
- approach: experimental assay
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (9 Nov 2020) · CC BY 4.0
clinker aligns translated genes from GenBank gene clusters, orders clusters by homology-plus-synteny similarity, and renders interactive, editable comparison figures via the D3-based clustermap.js library.
Comparing gene cluster homology across many loci is laborious. Existing tools (e.g. EasyFig, GeneSpy, genoPlotR) offer little interactivity, rigid styling, external BLAST steps or a need for programming, and clinker with clustermap.js aims to turn sequence files into publication-ready, to-scale comparison figures with minimal user effort.
Key results
Shownclinker aligns translated cluster genes against each other internally, so users do not need to generate BLAST comparison files themselves before plotting.
Global alignments via BioPython; defaults are BLOSUM62, gap open −10 and gap extension −0.5, editable through the clinker Python API. Hits under a user-set sequence identity cutoff are dropped. Multi-record GenBank inputs are treated as multi-locus clusters. · Gene cluster alignments using clinker
ShownThe default display order of clusters comes from a pairwise similarity score that combines sequence homology with conserved gene adjacency (adapted from MultiGeneBlast), followed by hierarchical clustering.
Score S = h + i · s, where h is the summed identity of homologous genes, s is the count of shared contiguous gene pairs, and the synteny weight i defaults to 0.5 (modified from Medema et al. 2013). Ward variance-minimization clustering runs in SciPy, and the dendrogram leaf order sets the display order. · Gene cluster alignments using clinker
Shownclustermap.js draws figures to scale and makes them editable in the browser: users can reorder, flip, trim, restyle or rename clusters, loci, genes, labels, links and legends, then export the figure as SVG.
Default scale is 15 pixels per 1000 base pair. Links are shaded by identity from 0% (white) to 100% (black), with a display threshold that can be raised or a best-hit-only mode. Single-linkage homology groups each get a unique colour, and clicking a gene moves its nearest homologs in other clusters into alignment. · Interactive visualisations using clustermap.js
SuggestedThe tool handles clusters split across contigs. The burnettramic acid biosynthetic gene cluster, fragmented over three contigs by the genome assembly, is shown with its homologues in one figure that needed only small external edits.
Figure 1b is reported as it appeared in clustermap.js; the only edit made in outside software was italicising species names. The static HTML for the figure is in the supplementary information. This is a single worked example with no quantitative evaluation. · Interactive visualisations using clustermap.js; Figure legends
Claimedclustermap.js works as a standalone library for any data that follows its schema. The authors present the pair as a downstream companion to cluster search tools such as MultiGeneBlast and cblaster.
The authors assert this without any benchmark against EasyFig, Artemis Comparison Tool, GeneSpy, Gene Graphics, GenomeDiagram or genoPlotR. No timing data or user-study data are reported. · Introduction; Interactive visualisations using clustermap.js
Methods, data and reproducibility
clinker parses GenBank files (multi-record files become multi-locus clusters), globally aligns translated genes with BioPython, scores cluster pairs by summed homology plus weighted synteny, and orders clusters by Ward hierarchical clustering. clustermap.js, built on D3, renders the result as an interactive, SVG-exportable figure.
- Burnettramic acid biosynthetic gene cluster (split over three contigs by fragmented assembly) and homologous clusters, used as the Figure 1b demonstration
- Methods & tools
- clinker, clustermap.js, BioPython (global pairwise alignment), BLOSUM62 substitution matrix, SciPy (Ward hierarchical clustering), D3 visualisation library, MultiGeneBlast (source of the cluster similarity formula, Medema et al. 2013), cblaster (upstream cluster search)
- Code
- http://github.com/gamcil/clinkerhttp://github.com/gamcil/clustermap.jshttp://pypi.org/project/clinker
- Software & environment
- clinker: Python 3, installable with pip from PyPI. clustermap.js: JavaScript built on the D3 library. Both are MIT licensed. No version pins or container are specified.
- Reproducibility gaps
- The text gives no package versions, no accessions or species for the example clusters, no runtimes and no test datasets. The static HTML behind Figure 1b is only in the supplementary information.
Limitations (2)
- No quantitative evaluation is reported: no runtime or scaling tests of pairwise global protein alignments as cluster numbers grow, no head-to-head comparison with EasyFig, GeneSpy or other cited tools, and the single showcase figure still received label edits in outside software. · noted in this summary
- Homology groups are built by single linkage of identity-thresholded links, which can chain distinct gene families into one group. The default synteny weight of 0.5 used for cluster ordering is not justified empirically. · noted in this summary
Read this if you need interactive, to-scale figures comparing homologous gene clusters (e.g. biosynthetic gene clusters) built directly from GenBank files, without writing plotting code or running BLAST separately.
- Citations
- 1,625
- FWCI
- 62.17
- Field percentile
- top 1%
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Evolclust: automated inference of evolutionary conserved gene clusters in eukaryotes
Marina Marcet‐Houben, Toni Gabaldón·BioinformaticsPeer-reviewed·Sep 26, 2019
- clusterSeq: methods for identifying co-expression in high-throughput sequencing data
Thomas J. Hardcastle, Irene Papatheodorou·bioRxiv·Sep 13, 2017
- cblaster: a remote search tool for rapid identification and visualisation of homologous gene clusters
Cameron L. M. Gilchrist, Thomas Booth, … Yit‐Heng Chooi·Bioinformatics AdvancesPeer-reviewed·Jan 1, 2021
- ClusterEnG: An interactive educational web resource for clustering big data
Mohith Manjunath, Yi Zhang, … Jun S. Song·PeerJ Computer SciencePeer-reviewed·May 21, 2018
- ExpressWeb: A Web application for clustering and visualization of expression data.
Bruno Savelli, Sylvain Picard, … Christophe Dunand·bioRxiv·May 2, 2019
- GALEON: A Comprehensive Bioinformatic Tool to Analyse and Visualise Gene Clusters in Complete Genomes
Vadim A. Pisarenco, Joel Vizueta, Julio Rozas·bioRxiv·Apr 15, 2024