- The shift
The next generation of biology runs on large, consistent datasets
Labs are automating to generate large, highly reproducible, high-fidelity datasets, both for screening and atlas projects and for the large models of cell biology that everyone is chasing. Instruments downstream already scale to thousands of samples. Sample prep is the step that still depends on hands and centrifuges, and that is where variability, cell loss and cost creep in.
Data quality
More than 95% viability, debris and dead cells removed, and native expression states preserved.
Throughput
96 samples per array: 5 minutes hands-on, 40 minutes of levitation, 45 minutes to enriched cells, several arrays per day.
Cost per sample
About $10 per sample instead of more than $100, with fewer failed libraries and repeats.
Automation
An SLAS/SBS footprint, scripted deck protocols and three hands-on steps.
- One automated workflow
Eight sample-prep steps, one platform
Enrich
Viable cells or nuclei levitate away from dead cells and debris.
Wash
Sequential buffer exchanges replace spin-and-resuspend cycles.
Pellet
Levitation gathers viable cells into a tight pellet above dead cells and debris, with no centrifuge.
Image
Integrated imaging records each sample as it levitates.
Analyze
Viability, debris and fraction metrics are calculated for every sample.
Collect
Only the fraction of interest goes forward.
Pool
Combine wells or replicates on the deck.
Concentrate
Elute in low volumes, ready for encapsulation, staining or plating.
- Any particle
Cells, nuclei, organoids and particles
Levitation separates by density and magnetic signature, not by labels. The same physics works across sample types, so one platform can serve many pipelines.
Cells
PBMCs, dissociated tumors, primary tissue, adherent and suspension lines, from human, non-human and plant sources.
Nuclei
Isolation, washing and elution in sequential wells for snRNA-seq and multiome.
Organoids and spheroids
Whole or dissociated organoids, standardized in 96-well format for screening and toxicity studies.
Particles
From bacteria to large cells and clusters. Debris and contaminants are tolerated and removed.
- Large models
Built for datasets that train large models
Virtual cell models, perturbation atlases and foundation-scale single-cell datasets learn from whatever you feed them. Dead cells, debris, ambient signal and operator-to-operator variability all become part of the model. Automated levitation removes those problems at the source.
- Reproducible inputs: identical, scripted processing in every well and every run limits batch effects
- Unbiased biology: no stains, beads, pressure or shear, so activation and expression states are maintained
- Rich metadata: imaging during levitation gives every sample a visual QC record and quantitative metrics to store with its downstream data
- Economics that scale: a lower cost per sample and fewer failed runs make million-cell studies practical
- For automation engineers
Integration without re-engineering
- Standard footprint: the SALT™ array is a 96-well SLAS/SBS ANSI plate format that sits in a normal deck position
- Simple deck layout: tips, levitation array and output plate block, with no centrifuge, vacuum manifold or magnet module
- Scripted programs: viable cell enrichment, cell washing and nuclei workflows run as protocols on the liquid handler
- Flexible inputs: 15,000 to 500,000 cells per well, with partial plates loaded row by row
- Scale-up path: develop methods on the LeviCell 12, then move the same chemistry to 96 wells on the LeviCell 96
The automated LevitasBio pipeline
LeviCell 96
The automated core: enrichment, washing and nuclei workflows for up to 96 samples on SBS robotics.
LeviSelect
Targeted depletion chemistry that runs alongside enrichment in the same wells.
LeviMetrics
The data layer: images and metrics from every levitation run, for QC and go/no-go decisions.
- Who automates with levitation
Built for teams generating data at scale
Pharma and biotech screening
- Organoid and primary cell dose response
- Cell-cell interaction and killing assays
- Pre-sort viability enrichment for cytometry
Genomics cores and service labs
- Pre-encapsulation viability enrichment for scRNA-seq
- Centrifuge-free washes for CITE-seq and CUT&Tag
- Nuclei prep for snRNA-seq and multiome
Atlas and large-model programs
- Thousands of samples processed the same way
- Per-sample QC images and metrics
- A lower cost per sample at every scale