Sequins: spike-in standards for RNA sequencing

[tedwong]

RNA sequencing (RNA-seq) has become a central tool with wide-ranging research and clinical applications. However, accurate resolution of gene expression is confounded by the sheer size and complexity of the transcriptome and further exacerbated by technical variables during library preparation, sequencing and bioinformatic analysis.

We have developed a suite of synthetic RNA isoforms (termed ‘sequins’) that align to artificial gene loci encoded within an accompanying in silico chromosome. Sequins can be spiked into a natural RNA sample and constitute internal controls for evaluation of the RNA-seq workflow, including library preparation, sequencing, split-read alignment, transcript assembly, gene expression, alternative splicing and fusion gene detection.

Please see our recent Nature Methods publication for a full description of sequins for RNA-seq.

This thread provides a forum for current and future sequin users to discuss any aspect of this emerging technology.

You can also watch this short video, which provides an introduction to sequins or visit http://www.sequin.xyz/ for more details.

Cheers

Simon Hardwick (on behalf of the Sequin team)

[dpryan]

Nice papers, I had already planned to talk about them in our internal journal club this week

[tedwong]

Thanks. Feel free to post any questions here. We will also release a software framework for the sequins very soon, and how sequins can be used in bioinformatics.

[dpryan]

Hi Ted, do you have any data regarding de novo transcriptome assembly (e.g., with trinity) of RNAseq data with sequins spike-ins? I imagine you could use the K562 data for that (at least from an initial read of the Hardwick et al. paper). We sometimes need to do RNAseq and transcriptome assembly of non-model organisms where people are explicitly interested in things like splicing. It'd be really nice to be able to use something like sequins to assess how much we might be missing (and assembly artefacts of course).

[dpryan]

Another question, how many reads were sequenced in the various experiments?

[tedwong]

Hi Devon, no we have not done any de novo assembly with Trinity. Our existing data could definitely be used to assess transcriptome reconstruction with Trinity, e.g. to measure sensitivity and precision of assembly (including an assessment of false-positives). Our data has been deposited with GEO (see paper for accession numbers) and will soon be available for download from our website at www.sequin.xyz.

For the K562 libraries (spiked with our RNA sequins Mix A), we sequenced ~70 million reads per replicate (for a total of ~210 million reads). For GM12878 (spiked with Mix B), we sequenced ~80 million reads per replicate (total ~240 million reads).

Cheers, Simon

[dpryan]

Thanks and congrats on the nice papers!

[tedwong]

Dear All,

We have now launched our website at www.sequin.xyz

Here, you can access useful resources and information on sequins. For those of you who are now starting to return sequencing libraries containing sequins, this is where you can find all the information you need, including:

1. Useful information - including design descriptions, protocols for using sequins in the laboratory, shipping information, and other handy documentation.

2. Anaquin Software – a software toolkit to analyse sequins. Anaquin is designed to integrate in standard NGS analytical pipelines, work with common bioinformatics tools, and supports standard data formats. A helpful User’s Guide provides detailed information and step-by-step tutorials.

3. Downloads – all the sequences, mixture files, annotations, example data files etc. you need to analyse sequins. These files can be used with Anaquin or with your own bioinformatic tools.

Currently the majority of information is focused on sequins in RNA sequencing, with further information on using sequins in metagenomics, whole genome and cancer genome to be released shortly.

Thanks for your support,
Sequin Team.