00 / Services
Research division

Elpis Research

What we build, and what you receive.

Six analysis domains, three engagement shapes, and a clear account of the work we don't take on.

01 / Domains
Where we move fast

Analysis domains.

Projects rarely sit neatly inside one row. These are the areas where the tooling already exists in-house.

Metagenomics & microbiome
shotgun · amplicon · MAG recovery · functional profiling
Taxonomic profiling from shotgun or amplicon data, de novo and reference-guided assembly, binning and metagenome-assembled genome recovery with quality assessment, functional and pathway annotation, and differential abundance testing with appropriate compositional handling.
Phylogenetics & comparative genomics
alignment · tree inference · orthology · selection
Multiple sequence alignment and tree inference at scales where standard tooling starts to break down, ancestral state reconstruction, orthology and synteny across genomes, and tests for selection. This is where the founder's open-source work sits, so unusual or large-scale requests are often tractable.
Variant analysis
WGS · WES · germline · somatic · population genetics
Germline and somatic variant calling from whole-genome or exome data, joint genotyping across cohorts, annotation and filtering designed around your actual question rather than a default threshold, and population-genetic summaries including structure, relatedness, and association testing.
Transcriptomics
bulk RNA-seq · single-cell · differential expression
Bulk RNA-seq quantification and differential expression with explicit attention to design and batch structure, single-cell processing from count matrices through clustering and cell-type annotation, integration across samples or modalities, and pathway-level interpretation.
Protein structure & function
structure prediction · structural comparison · variant effects
Structure prediction for proteins and complexes, structural alignment and clustering across families, pocket and interface characterization, and structure-informed interpretation of missense variants. Useful when sequence-level analysis has taken a question as far as it can go.
Pipeline engineering & reproducibility
Nextflow · Snakemake · containers · CI · rescue work
Converting an existing ad-hoc analysis — a folder of shell scripts, a notebook, an inherited pipeline nobody can run — into a workflow that executes reproducibly on a laptop, a cluster, or in the cloud. Includes containerization, dependency pinning, test data, and documentation.
02 / Shapes
Three engagement types

Engagement types.

Most projects fall into one of three shapes. All three are quoted as a fixed price after scoping.

Type

Analysis project

You have data and a question. We build the pipeline, run it, interpret the output, and hand over everything needed to reproduce it.

Best for a defined dataset with a defined question
Type

Pipeline build

You'll keep generating this kind of data. We build a reusable, tested workflow your team runs going forward, with training on handoff.

Best for recurring analyses and core facilities
Type

Consultation & review

Experimental design before you sequence, method review before submission, or a second opinion on an analysis you already have.

Best for grant prep, study design, manuscript revision
03 / Deliverables
What lands at handoff

What you receive.

  1. 01

    Results

    Processed outputs in standard formats, plus figures and tables prepared to publication standard where that's in scope.

  2. 02

    Source code

    The complete pipeline in a git repository you own, with version-pinned dependencies and container definitions. Not a black box, and not licensed back to us.

  3. 03

    Methods writeup

    Tools, versions, parameters, and rationale — drafted so it can be adapted directly for a manuscript or grant report.

  4. 04

    A walkthrough

    A working session with your team, so someone other than us can run the pipeline on new data before the engagement closes.

04 / Pricing
Quoted, not listed

Pricing.

We don't publish a rate card, because the same phrase can describe wildly different amounts of work. "Differential expression on 12 samples" is a few days. "Differential expression on 12 samples where the batches are confounded with condition and two libraries failed QC" is a different project entirely, and pretending otherwise at the quoting stage helps nobody.

Scoping conversations are free. Quotes are written, fixed-price, and itemized by deliverable. Academic and early-stage rates exist — ask.

05 / Limits
Work we decline

What we don't do.

Clinical or diagnostic analysis
not a CLIA or clinical laboratory
All work is For Research Use Only and must not be used to inform patient care or diagnostic decisions.
Wet lab work or sequencing
analysis only
We analyze data; we don't generate it. We're glad to advise on sequencing strategy and depth before you commit to a platform.
Guaranteed findings
methods scoped, conclusions never
If an analysis returns a null result, that is the result, and the writeup will say so plainly.
Analyses the data can't support
raised during scoping
If the design is underpowered or confounded for the question being asked, we say so before you pay — rather than delivering a technically valid pipeline that produces a meaningless number.

Start with a scoping conversation.

Describe your data and your question. If we're not the right fit, we'll say so — and where we can, point you somewhere better.

Request a quote