Lungfish Genome Explorer

Download. The current stable release is 2026.9.4. If you would rather try new features first, the preview is 2026.9.18. It changes more often and is more likely to surprise you. Both channels ship as signed, notarized disk images for Apple Silicon Macs running macOS 26 Tahoe or later.

Read the user manual, browse the source code, or see every release.

Introduction

I originally conceived of Lungfish Genome Explorer (LGE) as my love letter to Mac-based graphical sequence analysis software. Now I think it is more of a college-era mix tape: blending together a bunch of disparate things I like into what I hope is a thoughtful and opinionated whole.

LGE makes it easy for users to interact with many state-of-the-art open source bioinformatics tools that most users have to run from the command line. Most of the functionality is geared towards analysis and visualization of genomic data (Illumina, Pacific Biosciences, Oxford Nanopore) collected by labs and in NCBI SRA. LGE projects are designed for maximal reproducibility, can be shared easily with others, and can be extended by using agentic LLMs to perform analyses that are not built-in.

For example, if you have a bunch of FASTQ files and you want to map them to a reference using a tool that isn’t included, you can provide the path of the LGE project to Claude Code or Codex, and then have these tools run the mapping on data within the LGE project and save the results within LGE. I also encourage others to create their own LGE forks to meet their own individual needs.

Why did I make the Lungfish Genome Explorer?

I’ve spent my whole life using Apple computers, beginning with a Macintosh 512K my parents bought when I was in second or third grade. Knowing how to use Macs well was like a superpower growing up. I learned to type on Macs faster than I could write longhand, and made reports and newsletters, and, when I arrived at the University of Illinois Urbana-Champaign in 1994, took advantage of some of the earliest graphical internet tools on Macs including Eudora, TurboGopher, and NCSA Mosaic. Macs made it easy and fun to learn how to use the “information superhighway” and realize how much of the world’s information was suddenly at my fingertips.

When I started graduate school at the University of Wisconsin-Madison in 1997, my third rotation project involved generating and analyzing viral genomic sequences from animals infected with simian immunodeficiency viruses. This began a lifelong fascination with viruses, how they evolve to evade immune responses and cause disease, and how they move through time and space. Nearly 30 years later, I still get goosebumps when I analyze data on the viruses that are all around, and in some cases inside, us.

Fortunately for me, the lab where I did my Ph.D. exclusively used Macs. This meant that my first exposure to analyzing sequence data used graphical applications like Applied Biosystems’s AutoAssembler and MacVector. These tools taught me that the best way to understand viral sequencing data is to actually see it and interact with it. Within a few years, other applications that offered new tools for working with Sanger sequences and early deep sequencing datasets, like Sequencher and CodonCode Aligner, arrived and offered new capabilities. I used these tools every day for several years, before discovering Geneious (now Geneious Prime) in 2011. Geneious was a revelation. It was the first program that could work quickly with the virus genomic data we were generating on the Illumina MiSeq, and it could also handle data from the monkey genomics projects that were growing in size and complexity seemingly every day. It’s a testament to the quality of Geneious that I still have it on my laptop 15 years later.

Still, I occasionally became frustrated at some of Geneious’s limitations. One was cost. Geneious costs hundreds of dollars per year even for academic users. This isn’t unreasonable if you use it every day and have a well-funded lab in the US, but it can be out of reach for users who need to analyze sequence data infrequently, are just learning how to work with sequencing data, or are working on viral sequencing data from labs where this cost is unaffordable. There are some other free and open source graphical tools for working with sequencing data, including UGENE, but I personally haven’t found these to be intuitive.

Second, none of these tools are, as John Gruber would say, “Mac-assed Mac apps”. I’m writing this with BBEdit, a venerable text editor I’ve been using to write text for at least 20 years. It fits like a glove and looks like an application that was designed with care to use on a Mac. Existing graphical sequence analysis tools like Geneious and UGENE are cross-platform, which makes sense if the goal is the largest addressable audience. But this means that the features and aesthetics are reduced to the lowest common denominator supported by all operating systems.

Third, the ecosystem of bioinformatics tools is growing and evolving quickly. Geneious makes a curated set of these tools available in each release and has a plug-in system that allows technically inclined users to add new tools, but this is not straightforward. Workflow managers like Snakemake and Nextflow are not really supported. Graphical tools for working with sequence data often don’t trace the full provenance of analyses, making it difficult to know exactly what was done in each project. For example, if a graduate student gives me a Geneious folder, I often have a hard time understanding their chain-of-thought.

Fourth, the learning curve of these tools can be steep. Not as steep as using command line tools, but still a challenge for learners who are just beginning to work with sequencing data.

I have myriad other minor frustrations, as expected from anyone who has been working with any set of tools as long as I have. Until recently, there was nothing I could really do about these frustrations. I am grateful for the tools that exist. I run a busy lab. I might know what I want from a sequence analysis app, but I don’t have the programming expertise or the time to seriously take a run at making the Mac-assed Mac app that I’ve always wanted.

Lungfish Genome Explorer is unabashedly “vibe coded”

The advent of agentic coding changed this equation. For the first time, I could express in natural language what I wanted to create and have a system implement it on my behalf. In February 2026 I started building the Lungfish Genome Explorer (LGE) in Swift and AppKit, the native platforms for Mac apps. I didn’t, and still don’t, know how to write a line of code in these languages. However, I knew how I wanted LGE to behave, I’m obsessive, and I’m motivated to share the joy of working with genomic data with others. I’m also fortunate to be funded by Inkfish, which gives me precious time to be creative and try new things like work on LGE with the goal of making it freely available to scientists around the world.

About six months later, LGE is an application used by scientists in my lab every day. In another six months, I’m hoping that it will have largely replaced Geneious Prime. And I’m hoping that Dr. Heidi Horn will use it as the primary software tool used to teach undergraduate students in our UW-Madison class Pathology 501 how to work with viral sequence data.

One of the things I’m most excited about is how extensible LGE is. It can run tools in Docker containers, Conda environments, Snakemake or Nextflow workflows, or can be used “natively” with LLM tools. You can point an agentic LLM at an LGE project path and ask it to analyze the data within the project using tools that are not part of LGE. This makes it possible for novices to get their hands dirty with genomic data and gives power users the ability to do any sophisticated analysis they can imagine, as long as it uses tools that can be run on an Apple Silicon Mac or within a Docker container.

I’m also unabashed about hoping that others will make LGE their own by forking it and bolting on other features that they find useful. I expect that I’m going to be the primary maintainer of LGE and that I’ll continue developing it for my projects, but I also know that feature bloat is a very real risk. Instead of trying to take requests for new tools or workflows, I encourage you to create forks that you can tailor to your own needs by vibe coding the features you want it to have.

Where to go next

The user manual on Read the Docs is the place for step-by-step guidance. It is being written alongside the app, so expect gaps.

LGE is MIT licensed. Bug reports, failed workflows, and half-formed notes are all welcome on the issue tracker; please keep private sequence data and credentials out of public issues. I’m not taking pull requests, but if you fork it, the repository has the scripts to sign and publish your own build.

Development is supported by Inkfish.