Why we built sangeranalyseR: painless Sanger sequencing analysis in R

Sanger sequencing is the original way of reading DNA, and decades after it was invented it is still everywhere. It reads one fragment at a time and produces a chromatogram — the familiar four-colour trace of peaks — and it remains the everyday workhorse for confirming a cloned construct, barcoding a species, checking a CRISPR edit, or validating a variant a sequencer flagged. It’s accurate, cheap for small jobs, and woven into the daily routine of molecular labs around the world.

So here is something odd. As next-generation sequencing arrived, it brought with it a whole ecosystem of free, scriptable, interoperable software. Sanger sequencing — older, more widely used in many labs — never got the same treatment. The tools for it lagged behind.

sangeranalyseR is the package we built to fix that.

Why the old way was painful

When I was at the Australian National University, near the start of my research career, getting from a folder of .ab1 chromatogram files to a clean, aligned consensus sequence was more annoying than it had any right to be. Your options were essentially two. You could pay for commercial software — Geneious, Sequencher, CodonCode Aligner — which is capable but expensive and awkward to fit into an automated pipeline. Or you could wrestle with venerable command-line tools like Phred, Phrap, and Consed, which are free but cumbersome and hard to glue to anything modern.

What there wasn’t, conspicuously, was a good option in R — the environment where so much of bioinformatics and phylogenetics actually happens. There was one R package, sangerseqR, but it worked on a single read at a time; it didn’t do the real, repetitive job, which is assembling many reads into contigs and aligning those contigs across samples. So, with Rob Lanfear and colleagues, I built the tool I wished I’d had.

What sangeranalyseR is

sangeranalyseR is a free, open-source R/Bioconductor package that carries you all the way from raw chromatograms to aligned consensus sequences. It leans on sensible defaults so the common case is trivial, while exposing every parameter for when you need control.

⤢ Enlarge

✕

The structure mirrors how Sanger data is actually organized: individual reads (SangerRead) assemble into per-sample contigs (SangerContig), which align into a final multiple alignment (SangerAlignment). And the headline claim is real — a complete analysis is about four lines of R:

library(sangeranalyseR)

# one call: load every .ab1 file, quality-trim, assemble contigs, and align them
aln <- SangerAlignment(inputSource = "ABIF",
                       parentDirectory = "./Allolobophora_chlorotica")

writeFasta(aln)       # write the aligned consensus contigs to FASTA
generateReport(aln)   # write a full, interactive HTML report

What it does

Four lines, but flexible underneath. That single SangerAlignment call does a lot: it reads every chromatogram, trims each one by quality (using either the modified-Mott algorithm from Phred or a sliding window in the style of Trimmomatic), assembles the forward and reverse reads of each sample into a consensus contig, and aligns the contigs together. The figure below shows it on a real dataset — eight earthworm (Allolobophora chlorotica) samples from the Barcode of Life Database, sixteen .ab1 files in all.

⤢ Enlarge

✕

Interactive when you want to look closer. Good defaults handle most reads, but some need a human eye. So sangeranalyseR ships with interactive Shiny apps that let you inspect and adjust everything without leaving R: view each read’s chromatogram with the trimmed ends shaded, see per-base quality scores and secondary peaks, scan a heatmap of pairwise differences between the reads in a contig, and — if you supply a reference protein — catch indels and premature stop codons. You can re-trim a read by dragging a slider, watch the consensus update, and then save those exact parameters so the whole analysis is reproducible.

⤢ Enlarge

✕

And it lives in R. Results come back both as standard FASTA files, for handing to other tools, and as R objects, for the rest of your analysis. The package is MIT-licensed, distributed through Bioconductor, documented in depth, and unit-tested — the kind of plumbing that makes a tool dependable rather than just clever.

The bigger picture

Not every useful contribution is a new algorithm. Sometimes it’s meeting a widely used but underserved method where its users already are, with good defaults and a gentle on-ramp. Sanger sequencing isn’t fashionable, but it’s in nearly every molecular lab, and making its analysis free, scriptable, and reproducible quietly removes friction for a lot of people.

This was my first real software project, and it set a pattern I’ve followed ever since: when the tool you need doesn’t exist, build it well, document it, give it away, and use it yourself. Build what you need, use what you build.

---

Read the paper in Genome Biology and Evolution, install it from Bioconductor, browse the code, or work through the documentation. sangeranalyseR was built with Kirston Barton, Sarah Palmer, and Robert Lanfear.

Further reading: sangerseqR, the R package for handling individual Sanger reads that sangeranalyseR builds on and extends.