seqsalvage reads data from proprietary molecular biology software into R. It currently supports Sequencher project files (.SPF). It can extract the chromatograms, base calls, quality scores and peak positions stored in a project, and write them back out as .ab1 trace files.
We aim to support open science and reproducible research by reducing dependence on proprietary software, which can lock a lab out of its own data once the software stops running. Sequencher 5.4 for Mac, for example, is a 32-bit application and no longer opens on current macOS. Since these file formats are not publicly documented, the parsers are developed through reverse engineering and implemented directly in base R. If you run into a file that doesn’t parse correctly (or at all), please open an issue.
Usage
library(seqsalvage)
reads <- read_spf("project.SPF")
reads[[1]]$name
substr(reads[[1]]$bases, 1, 60)
matplot(reads[[1]]$traces[1:600, ], type = "l", lty = 1, col = c("green3", "blue", "black", "red"))
# write every read in the project as an .ab1 file
spf_to_abif("project.SPF", "recovered_traces")
# also fill the base-caller tags, so Biopython and sangerseqR can read the files
spf_to_abif("project.SPF", "recovered_traces", basecaller = TRUE)
# include reads deleted from the project whose data is still in the file
deleted <- Filter(function(r) r$deleted, read_spf("project.SPF", deleted = TRUE))
# contigs: name, saved consensus and member reads
contigs <- read_spf_contigs("project.SPF")
contigs[[1]]$name
contigs[[1]]$reads
# write each contig's saved consensus to a FASTA file
spf_contigs_to_fasta("project.SPF", "contigs.fasta")Each read comes back with its file name, base calls (including any edits made in Sequencher), quality scores, peak positions and the four processed trace channels. A project keeps only the edited calls and doesn’t mark which reads were edited, so spf_to_abif() writes the calls to the ABIF tags for user-edited calls. Biopython and sangerseqR read only the base-caller tags; basecaller = TRUE copies the calls there too, at the cost of labelling any edits as base-caller output.
Status
The Sequencher parser was reverse engineered from a project saved by Sequencher 5.4 on a Mac. For the 150 reads in that project whose original .ab1 files survive, it reproduces the processed trace channels, peak positions and quality scores exactly. It does not recover the raw channels, so the .ab1 files that spf_to_abif() writes are not copies of the originals. read_spf() also reads all 2,213 reads in the 14 Sequencher projects published with Claussin and Chang (2016) and Zenodo record 5557681, though these come without original files to compare against.
read_spf_contigs() reads each contig’s name, saved consensus and member reads, but not where each read sits in the contig. In the published projects, it places every read in a contig. ?spf_format describes the format.