The Platform

One guided cycle takes sequence data to ranked strain candidates.

StrainX Works combines genome-scale metabolic models, generative sequence proposals, and fermentation constraint scoring in one iteration workflow.

One workflow, start to finish.

Three nodes: data in, candidates out, each step tied to a predicted pathway outcome.

Data Input Generative Engine Ranked Candidates

Your inputs

Bring data in, build the model.

The platform takes the data already held by your fermentation team. No special preparation or proprietary-schema reformatting.

Genome assembly, reference sequence
Target molecule identity, target titer, and yield specification
Fermentation media composition, bioreactor scale parameters
Historical growth curves and metabolite data (optional, model quality improves)

Supported formats

FASTA genome sequence files
FASTQ short-read data
SBML metabolic model files
JSON metabolic model definitions
CSV fermentation run logs
XLSX fermentation data exports

Standard ELN exports accepted. Proprietary LIMS API integration is on the roadmap.

Engine

For non-model organisms.

Most platforms assume well-characterized chassis strains. StrainX Works builds an organism-specific model from your data, making novel or under-characterized production strains common in industrial biomanufacturing viable. Each project starts with your organism's actual genomic and fermentation data, rather than a generic reference.

The result is a model tuned to your strain background, fermentation conditions, and target molecule, not a statistical average of published literature strains.

Graph-based metabolic representation
+ transformer-style sequence scoring.

Predicts flux impact of gene deletions,
insertions, and promoter swaps across
your specific strain background.

Model improves with each validation
cycle. Iteration 1 is informed by
your input data. Iteration 2+ is
informed by your lab results.

Platform result

Ranked candidates, not searches.

Every iteration cycle yields a prioritized set of strain modification candidates, scored against your titer target, tolerance constraints, and fermentation scale. These are not literature hits or generic optimization suggestions. They are candidates ranked for your organism, molecule, and fermentation environment.

Illustrative citric acid titer optimization output (synthetic data for illustration)

Rank	Modification	Yield Delta	Tolerance Score	Priority
01	citB-P37L + oadA-OE + sdhB-KO	+34%	0.87	HIGH
02	citB-P37L + mcsA-A114V	+28%	0.91	HIGH
03	oadA-OE + icd-G407R + fumC-KO	+21%	0.78	MED

Propose Validate Refine

Cycle loop

The model learns from each lab cycle.

Lab validation results flow back into the model. Better-than-predicted candidates revise the flux assumptions; underperforming candidates reveal constraints not yet modeled. Proposals improve with every iteration cycle.

2-3 cycles to a production-ready strain, most partners
2-3 weeks average per cycle
8-10 weeks typical from project start to production-ready strain

Compatible data.

Takes the file formats your lab already uses. No conversion pipeline is needed.

Genome Sequence

FASTA
FASTQ
GenBank flat files

Metabolic Models

SBML (.xml)
JSON (COBRA)
MAT (legacy)

Fermentation Data

CSV run logs
Excel exports
Standard ELN formats

API (Roadmap)

LIMS integration (planned)
ELN webhooks (planned)
REST API beta (Scale tier)

Ready to shorten your strain development timeline?

Join the early-access program. Our team works directly with new partners on the first iteration cycle.

Request Access Talk to a Scientist