I am releasing kyber today, and you can find all about it on GitHub. In short, kyber produces a heatmap showing where the majority of your reads are. On the x-axis is the log transformed read length, on the y-axis the accuracy (optionally Phred-scaled). The axis ranges are fixed and therefore easier to compare various… Continue reading Introducing kyber: minimalistic quality assessment plot for long reads
Long-read sequencing enables the phasing of variants and reads, i.e., separating those into two parental haplotypes. Without trio information, you will not be able to say which variant is inherited from which parent, but you will know which variants were inherited together. And this matters, e.g., for compound heterozygous variants and cis-regulatory variation. Phasing can… Continue reading Announcing Phasius for visualization of phase blocks
Of course, having a newer version of the nanopore base caller will have some impact. The accuracy obviously constantly improves, and in long-running projects like ours, we now have samples with Guppy4, Guppy 5, and Guppy 6 base caller versions.I reasoned that it would probably be a confounder, but we mainly look at structural variation.… Continue reading Beware of confounding basecaller versions
I am releasing Cramino today, my first Rust experience. Cramino is a faster replacement for NanoStat, and extracts features from cram and bam files that are useful for quality assessment of long read sequencing data, including read lengths and read identities. The default output looks like below (left), and is generated in 12 minutes for… Continue reading Introducing Cramino: a *fast* QC tool
A (for us) very useful diagnostic of a nanopore run is the status of the pores: if they’re sequencing (‘single_pore’), saturated, unavailable or multiple. It is also interesting to see how fast you are losing pores or killing your flow cell with e.g. a particularly blocky library. Fortunately these metrics are since not-too-long ago saved… Continue reading Comparing the pore status of flow cells
I’m a big fan of saving time by automation. Any job you have to do repeatedly which doesn’t require your brain would be a job to be automated. Below is how I organize my workflow and the tools which I use to stay up-to-date with the scientific literature and to organize my reading. I use… Continue reading Automate everything: following and reading literature
In surpyvor I collect scripts which can be helpful for structural variant analysis. Today I added a “purge2d” subcommand to remove accidental 2D/1D^2 nanopore reads from a BAM file. Most typically we do 1D ligation preps, but in some cases the complement fragment is sequenced directly after the template and not recognized as separate molecules,… Continue reading Removing accidental 2D/1D^2 reads from an alignment
We recently published methplotlib, a tool for the visualization and analysis of modified nucleotides from nanopore sequencing. It works downstream of tools like nanopolish, nanocompore and direct methylation calling by the guppy basecaller. More information can be found on GitHub. Feedback, suggestions, reporting problems and feature requests are very much appreciated. Below are some example… Continue reading Methplotlib examples
Since a recent version of MinKNOW, the software controlling a nanopore sequencer, a sequencing_summary file is created before basecalling, in which one column is of particular interest: the end_reason. Although I’m not yet sure what each value means, I believe it gives per read the reason why the software decided to stop sequencing here, mostly… Continue reading Comparing the end reason of reads in a nanopore experiment
I wanted to make a stacked bar chart to show the number of variants with a certain FILTER status per sample from a multi-sample VCF. Nowadays I make all plots with plotly, because it’s fast, convenient to write and dynamic HTML makes it easier afterwards to select the bits I’m interested in to show. In… Continue reading Stacked bar chart of FILTER information from a multi-sample VCF
Gigabase or gigabyte 