Discovery & pre-clinical

Early-stage work rewards breadth and speed. We quantify large image sets in detail so you can compare compounds, doses and models on numbers rather than impressions — and kill the weak candidates before they consume a study budget.

What this service answers

At discovery you are usually asking one of three questions: is the target where I think it is, did the compound do anything, and does the effect scale with dose. All three are measurable in tissue if the quantification is careful enough.

Target expression and localisation — quantification of functional biomarkers on a cell-by-cell basis, including whether signal sits in the nucleus, membrane or cytoplasm.
Effector cell identification — accurate detection of specific populations in diseased tissue, such as CD8-positive or regulatory T cells within a tumour.
Tissue invasiveness — cell numbers and densities at the tumour–stroma border, and at defined distances either side of it.
Compound efficacy — drug response profiled directly in the affected tissue rather than inferred from a downstream surrogate.
Dose relationships — consistent quantification across dose groups so that trends can be tested statistically.

Screening scale without losing cell-level detail

A single algorithm, developed on a representative subset and then applied across the full image set, gives you a comprehensive high-throughput screen where every slide has been measured the same way. Manual scoring cannot do this at scale, and it cannot do it without drift.

Typical study size

40–2,000 slides

Feasibility turnaround

5–10 days

Markers per panel

1–12

Output granularity

Per cell, per region, per slide

What lands in your inbox

Object-level data
One row per detected cell, with position, classification, marker intensities and the region it fell in. Delivered as CSV or Parquet.
Summary tables
Aggregated by slide, animal, group and region of interest, with the grouping variables you specified in the statement of work.
Mark-up images
Overlays showing exactly what the algorithm classified as what, so your pathologist can sanity-check the output rather than trust it.
Methods report
A written description of annotation strategy, algorithm design, thresholds and exclusions — detailed enough to reproduce or to drop into a publication methods section.

Next phase: clinical How the models are built

Send a pilot set first.

Ten representative slides usually tell us whether your endpoint is measurable, what the staining will allow, and what the full study should cost.

Start a study Browse services