A brief history of Ziplign

If you are interested in why Ziplign was made in the first place, and some of the details of the implementation, then please read on.

Motivation

Artemis and ACT are excellent programs, both of which I use heavily. Sadly, they are no longer really actively maintained. There are lots of great genome viewers out there - obviously IGV being very popular. Implementing yet another genome viewer is probably not a good use of time.

However, there is no tool for me that can do what ACT does. It is extremely comprehensive (because it essentially has the functionality of Artemis). But there are features that I always wish it had:

  • proper support for multi-FASTA files (you can hack to get around this)

  • run BLAST for you, instead of needing to provide your own comparison file

  • show base-level alignment between genomes (SNPs, indels). I have always wanted to see these when zoomed in to the sequence level.

  • easier contig reordering, and also be able to reverse complement individual contigs

  • out of the box support for Apple silicon. It does run on Apple silicon, but needs a bit of Java-related hacking to do so. This seems beyond the skills of less technical users

  • save and load a “project” - ie both genomes and the comparison file - instead of needing to keep 3 files around for each use of ACT

I was playing around with the game engine Godot, and tested how difficult it would be to get the basic functionality of ACT working. This was just showing rectangles (contigs), parallelograms/triangles (blast matches), and being able to slide the top and bottom genomes around and have the blast matches follow suit. It turned out to work quite well as a proof of concept. This then grew into Ziplign.

The aim of Ziplign was always to be a minimal implementation of the essential things I want from ACT and nothing more: view two genomes, plus the features listed above. It’s pronounced “zipline” and is a backronym: “Zoomable Interactive Paired-genome aLIGnment Navigator” (or have fun making up your own acronym). I rarely use ACT to view more than two genomes, so early on I decided (to save a lot of implementation pain) to only support two genomes. This is baked in and will not change.

Under the hood

I wanted Ziplign to be cross-platform, which was another reason to use Godot. It can export a project to Windows, Mac, and Linux (and x86 and ARM architecture). It can also make Android and i(pad)OS apps - which I thought about making, but I don’t think it’s possible to get them to run BLAST. It would have been nice to have an ipad app, but it is too locked down to make this practical.

This was my first project using Godot (and indeed a GUI of any kind). Handling all the 2D graphics in Godot is relatively easy, what with it being a game engine. Displaying genomes and BLAST matches was probably the simplest part. The biggest challenge of making Ziplign was the GUI: adding menus, buttons, and then in particular being able to change colour themes etc. Most of this comes down to Ziplign not being a game, but a tool. Although Godot is set up to make such a program (its own IDE is made in Godot), it is primarily for making games.

There is, unsurprisingly, no “bio-Godot”. Seeing as Ziplign would need to run BLAST as a separate program, I decided to offload all the heavy lifting bioinformatics tasks to a separate program as well, called zlhelper. This also needed to be cross-platform, and so is written in the Go programming language. It parses sequence files, handles reading compressed files, runs BLAST and parses its output. The plus side of using Go is that it can compile stand-alone binaries for Windows, macOS and Linux, and both ARM and x86 architecture.