Platform basics
What is Digital Seed?
Digital Seed is a cloud-based plant genomics platform that turns marker and phenotype data into breeding decisions. It bundles GWAS, genomic selection, population structure, multi-environment trial analysis, selection indices, gene annotation and environmental intelligence into a single web workflow — no server setup, no scripting required.
Who is it for?
Public breeding programs, seed companies, agri-biotech startups, and plant science researchers. If you have SNP genotypes and phenotype records for a crop, you can use Digital Seed.
Do I need to install anything?
No. Digital Seed runs entirely in your browser. All computation happens on our managed backend.
Which crops are supported?
Species-agnostic — any diploid or polyploid crop with SNP data works. We ship curated reference panels and gene models for major crops (wheat, maize, rice, soybean, tomato, cotton and more); see the Crops page for the full list.
Data & file formats
What genotype formats can I upload?
VCF (v4.1+), PLINK (.bed/.bim/.fam), HapMap and tab-separated numeric matrices. Files can be gzipped. Multi-sample VCFs up to several GB are supported.
What phenotype formats work?
CSV or TSV with one row per genotype/environment. The first column is the accession ID; subsequent columns are trait values or factor levels. Missing values as NA or empty are handled automatically.
How long is my data stored?
By default, all uploaded datasets, analyses and results are automatically deleted 20 days after upload. Export any results you want to keep. Paid tiers can extend retention.
Is my data shared with anyone?
No. Your data is private to your account and protected by row-level security. We do not train models on your data and never share it with third parties. See the Privacy Policy for details.
Methods & results
Which GWAS models are available?
Mixed-linear model (MLM), compressed MLM, FarmCPU and BLINK, with optional kinship and principal-component covariates. Multiple-testing correction uses Bonferroni or Benjamini–Hochberg FDR.
How is genomic selection implemented?
GBLUP by default, with optional Bayesian ridge, BayesA, BayesB and RKHS models. Cross-validation splits, prediction accuracy and reliability are reported for every run.
Can I reproduce a run?
Yes. Every analysis stores its exact inputs, parameters, software versions and random seed. Any run can be rerun with one click, and results can be exported as BibTeX or RIS for citation.
How long does a typical run take?
A GWAS on 500 accessions × 50k SNPs finishes in 1–3 minutes. Genomic selection with 5-fold cross-validation on similar data takes 3–8 minutes. Progress is streamed live.
Accounts, pricing & quotas
Is there a free tier?
Yes. The free tier includes a monthly quota of analyses, limited storage and a cap on concurrent runs. See the Pricing page for current limits.
What happens if I hit my quota?
New runs are blocked until the next monthly reset or until you upgrade. Existing results remain accessible. Quota usage is visible in your dashboard at any time.
Can I cancel anytime?
Yes. Cancellation is self-service from your account settings and takes effect at the end of the current billing period.
Security
How are passwords protected?
Passwords are hashed with bcrypt and checked against the HaveIBeenPwned leaked-password database at signup and reset. We support email OTP and OAuth sign-in.
Do you offer SSO?
Google OAuth is enabled by default. SAML SSO for institutions is available on the enterprise tier.
Where is data hosted?
On managed cloud infrastructure with encryption in transit (TLS 1.3) and at rest (AES-256). Backups are geographically redundant.