Discovery Engine

++ THE ENGINE, OPENED
TECHNOLOGY 01 MatrixFOREST COMPREHENSIVE RNA-⁠SMALL MOLECULE BINDING SYSTEM TECHNOLOGY 02 SpliceVerse COMPREHENSIVE SPLICING CASSETTE-⁠SMALL MOLECULE ASSESSMENT SYSTEM TECHNOLOGY 03 SAR by Selectivity SELECTIVITY AS THE METRIC
TECHNOLOGY 01 / 03

MatrixFOREST

The technology that comprehensively captures the compounds that bind RNA. An RNA structure library and a compound library are analyzed against each other in one experiment, so affinity, selectivity and binding site arrive together rather than from three separate experiments.

The MatrixFOREST profiling matrix One assay drawn as a three-dimensional data volume, seen isometrically. The floor is a compound plate ruled into ~1,000 fine wells, standing for a 100K+ compound library across both of its axes; a turquoise band sweeps the plate and weak binders light up in waves. Thin beams of light rise vertically out of the wells that reacted. Stacked above the plate are translucent planes, one per RNA structure in the library, fading upward to say that 1,000+ of them are there. A horizontal reading plane travels up the stack, reading one RNA plane at a time. Where a beam meets one plane a single voxel ignites in gold — one selective hit inside the volume — announced by elliptical sonar rings lying in that plane and traced out to a readout panel where a compound-free control trace and a compound-added trace are overlaid; the binding site is called where the compound-added trace rises above the control, and it is marked on the RNA fold. RNA STRUCTURE LIBRARY 1 PLANE = 1 RNA · ×1,000+ READOUT MUTATION RATE BINDING SITE 100K+ COMPOUNDS ~109 PAIRS · READ IN ONE ASSAY SELECTIVE HIT WEAK BINDER
Schematic of the library-versus-library volume and its readout. The plate face is the compound library, the vertical axis is the RNA structure library, and a hit is one voxel inside the solid. Binding sites are called from the difference between the compound-free control and the compound-added trace. Positions and traces are illustrative, not measured data.
INPUT
RNA structure library × compound library
OUTPUT
Affinity for RNA, selectivity across RNA structures and the compound's RNA binding site
ROLE
Mapping the RNA-binding chemical space, and obtaining hit compounds
RNA STRUCTURES / TUBE1,000+
INTERACTIONS / RUN~109
CHEMICAL LIBRARY100,000+
TECHNOLOGY 02 / 03

SpliceVerse

The technology that comprehensively captures the compounds that control splicing. A splicing cassette library and a compound library are evaluated in one experiment, so the strength of splicing control, selectivity across cassettes and the type of splicing change induced all arrive together.

Compound-dependent splicing outcome A pre-messenger RNA with three exons and two introns. A compound acts on the middle target exon, tipping the junction choice. Above the strand, turquoise arcs show the two options being read. Below right, two transcripts are read out in gold: an up-triangle outcome where the exon is included and up-regulated, and a down-triangle outcome where a poison exon is included and the message is down-regulated. At lower left the cassette library is drawn as a three-dimensional volume in the same isometric language: a finely ruled compound plate on the floor, translucent planes stacked vertically above it — one per splicing cassette — fading upward toward 1,000+, beams of light rising out of the wells that responded, and a reading plane travelling up the stack. One plane is framed in gold and a leader line carries it up to the strand above: the mechanism drawn here is one cassette out of that stack, evaluated against 100K+ compounds in one reaction. COMPOUND PRE-mRNA · TARGET EXON READOUT — SPLICING OUTCOME · MoA 1,000+ CASSETTES UP · EXON INCLUSION DOWN · POISON EXON 1,000+ CASSETTES × 100K+ COMPOUNDS ~109 EVALUATIONS · READ IN ONE ASSAY
Schematic of a compound-dependent splicing outcome. The lower-left volume is the cassette library read in parallel — the plate face is the compound library, the vertical axis the cassettes — and the strand above is the one plane framed in gold. Illustrative, not measured data.
INPUT
Splicing cassette library × compound library
OUTPUT
Splicing activity, splicing selectivity and splicing type
ROLE
Mechanism-led hit discovery and selectivity profiling
CASSETTES / WELL1,000+
EVALUATIONS / RUN~109
CHEMICAL LIBRARY100,000+
TECHNOLOGY 03 / 03

SAR by Selectivity

For each target we design a bespoke mini-library of splicing cassettes, putting the on-target and its main off-targets on a single panel. That lets us quantify both at the same time and run structure-activity relationships (SAR) weekly, with selectivity across multiple splice sites as the metric. This same selectivity engineering also lets us convert a non-selective compound onto a different target, widening the therapeutic window and bringing previously undruggable targets within reach.

The weekly selectivity cycle and target conversion Upper register: three stations arranged in a loop — a SpliceVerse profiling assay, a selectivity-evaluation heatmap, and compound design — joined by a return arrow that closes the cycle, annotated as one week per turn repeated n times, ending in an improved-selectivity compound. Lower register: one non-selective splice modulator branches through target conversion into two compounds, each selective for a distinct target marked in gold. SPLICEVERSE SELECTIVITY DESIGN 1-WEEK CYCLE  ×n NON-SELECTIVE ON-TARGET NEW TARGET TARGET CONVERSION
Schematic of the weekly selectivity-SAR cycle and of target conversion. Loop stages, cell positions and target marks are illustrative, not measured data. Gold marks a reached selective target; the on-target branch widens the therapeutic window, the redirected branch reaches a previously undruggable target.
INPUT
Selectivity measured at the first screen, carried into each design round
OUTPUT
A selectivity-qualified lead series
ROLE
Hit-to-Lead optimization on weekly selectivity-SAR cycles
POTENCYMagnitude of each type of splicing change
SELECTIVITYSplicing changes across 100+ cassettes
CYCLEWeekly

TECHNOLOGY 01 / 03 — MatrixFOREST

++ THE VERIFICATION LAYER

Nothing leaves the engine on one measurement

The three technologies propose. Two further capabilities test what they propose, in cells and by an orthogonal route. They are not a fourth station on the conduit; they run beneath it.

In-cell RNA structure probing

RNA structure is measured directly in cells, to test whether a compound actually changes the structural neighborhood it was selected for.

INPUT
Live-cell RNA structure × candidate compound
OUTPUT
High-resolution, compound-dependent structural change
ROLE
Target engagement and cellular mechanism validation

RNA-specialized analytics

RNA-focused assays give an independent line of evidence for candidate interactions, so an SAR decision never rests on a single method.

INPUT
Candidate RNA–compound pairs
OUTPUT
RNA-focused affinity and selectivity measurements
ROLE
Orthogonal validation and SAR decision support
++ WHERE IT LANDS
Our Target & MoA

Why structured RNA is addressable, and how small-molecule engagement can redirect RNA fate.

OPEN → The pipeline

Six programs, and how far each has come. Target, mechanism and stage are disclosed only where a partner agreement allows it.

OPEN → Publications

The peer-reviewed papers behind the methods on this page.

OPEN →

← BACK TO THE TECHNOLOGY OVERVIEW