Your Data Has More Than One Story

“I have a biological question. I don’t yet know the best way to answer it.”

Study Design · Sequencing Strategy · Statistical Power

The decisions made before a single sample is sequenced determine what your data can and cannot tell you – and what it silently cannot.

The wrong design doesn’t announce itself. It quietly wastes a season of work and leaves you with data that cannot answer the question you actually had. We work out what depth, what replication and what design will answer your question before anything is sequenced – and what will silently fail to. This is where an underpowered study gets caught before it becomes one.

Design isn’t only “should we sequence.” It’s also “are we sequencing the right thing” – targeted panels instead of whole-genome when only a defined set of genes matters, the difference between a design that answers your commercial question and one that just generates data.

Where this fits: upstream of everything below. Get this step right and the rest gets easier.

“I have strains, samples or sequencing data. I know what I’m looking for – finding it is the bottleneck.”

Genomics · Transcriptomics · Microbiomes · Metagenomics · Protein Function

Core analysis across whichever layer your question lives in – from genome assembly and variant calling to gene expression, microbiome composition and protein function.

Genetic variation and genome assembly. We benchmarked sequencing and optical mapping technologies head-to-head for plant genome assembly, in a study still widely cited by genome projects choosing a sequencing strategy (GigaScience, 2019). We compared 1,524 pathogen genomes to resolve pathogenicity factors in pangenomes (Cell Host & Microbe, 2018), and showed that across 187 wild plant species, a larger immune gene repertoire comes at a direct cost to growth – a trade-off largely absent in domesticated crops (Science, 2024).

Protein structure and binder design. We model protein structure, predict the functional impact of sequence variants and support epitope mapping and binder design for biotech clients under NDA – work that has fed directly into live antibody and phage engineering pipelines.

Gene expression. We built and published a spatial transcriptomics method that reads tissue-by-tissue gene activity at a fraction of standard cost (Plant Methods, 2019). A gene present is not a gene active. We find the difference.

Microbiomes and metagenomes. We work at strain-level resolution throughout – because two “similar” microbiomes can mean opposite outcomes for your host. We developed our own extraction methods for recovering microbiomes directly from plant tissue and soil, contributed to AirSeq, a portable airborne eDNA method that detects crop pathogens and pests directly from field air samples (Current Biology, 2024), and traced how plant genetic background shapes microbial community composition across soil adaptation gradients (preprint, 2025).

We think across species and environments. The analytical and methodological approach transfers, even when the biology varies. That’s the pattern we bring to every project.

“I have results from several angles, but not the mechanism that connects them.”

Multi-Omics Integration

Separate findings from separate analyses become a single connected explanation – the mechanism that links them, not three results sitting side by side.

A variant, an expression change and a microbiome shift can each be real and still tell you nothing on their own. Connecting them means knowing which links are meaningful and which are coincidental, and being able to show the reasoning either way. That is what integration is for: not more data, but the right interpretation of the data you already have.

“I need this to hold up – in review, in a board meeting, in an audit.”

Documentation · Reproducibility · Manuscript & Diligence Support

Every project delivered with version-controlled code, documented decisions and publication-quality figures – the same standard that gets work through peer review at Science, Current Biology and PNAS.

The bar we hold ourselves to is the same one that got our own work through review at Science, Current Biology, Nature Communications, Cell Host & Microbe and PNAS – work judged by independent referees with every incentive to find a flaw, and didn’t. That standard travels with every project, whether the destination is a manuscript, a regulatory submission or an internal R&D decision.

For biotech clients, that means work built to survive due diligence: methods a different team can pick up and reproduce, decisions documented well enough to defend to a board member who is also a scientist and an audit trail that holds regardless of who is asking.

“I don’t have one defined project. I need a hand that’s just there – or I need to win the funding first.”

Ongoing Support · Single Analyses · Grant Development

A single analysis, a standing arrangement or grant support including preliminary data and methods – including before there is a funded project to talk about.

Not every engagement starts with a fully scoped question. Some clients want a standing arrangement to call on as things come up – capacity during a funding round, a regulatory submission or while a permanent hire is being recruited. Some want exactly one analysis done properly, with no commitment beyond it. And some don’t have a project yet – they need the preliminary data and analysis plan that makes a grant application credible. We have been through that review process ourselves. We know what scrutiny looks like from the other side.