01/splicify/core
Annotation engine, clean-room
MIT-licensed replacement for pLannotate. Six reference tiers searched in parallel — GenoLIB, FPbase, curated SwissProt, curated Rfam, RefSeq, and a motif DB.
~3.8s14.8 kb lentiCRISPR v2
◆ v0.1 / annotation + cloning
A plasmid is
a validated
functional circuit.
For 15 years plasmid tools gave us flat feature lists. Splicify · Clone emits an interactive module graph representation, chooses your parts and workflow, clones your plasmid in silico, and validates your product’s internal grammar.
›
upload inventory for reference · optionally add a target for a specific end point
Step 01 / 05
00%
Parsing sequence
Reading GenBank · detecting topology
···
Annotating features
6 reference tiers · feature type and function lookup
Scoring workflows
6 cloning strategies · feasibility ranked
Assembling in silico
Picking least-work, highest-success route
Validating product
Sequence + module + interaction invariants
/ What shipped
6 modules · all first-class
02/splicify/core
Hierarchical module detection
Not features — architecture. LTR-bounded lentiviral payloads. T-DNA LB/RB. Gateway quartets. EBV oriP + EBNA1. OsTIR1+AID degrons.
~80rules across 30 categories
03/splicify/core
SBO-typed interactions
Every plasmid emits a graph of genetic regulation — promoter → CDS [NLS · Cas9 · P2A · PuroR] → Poly A signal. A plasmid has grammar rules and interactions.
graphfirst-class design artifact
04/splicify/core
Cloning-feature scanner
Type II + Type IIs sites strand-aware. All 22 canonical Gateway att cores classified. PCR feasibility flags for GC extremes, repeats, palindromes.
~0.09sper plasmid
05/splicify/core
Target-from-Inventory auto-routing
Upload target + freezer inventory. We score ten cloning workflows, pick the method with the least hands-on work and highest wet-lab success, and hand off to the right designer.
10workflows scored in parallel
06/splicify/core
Plasmid language model
Build plasmids de novo and create synthetic fragments that align with your description
closedgeneration ↔ validation loop
/ Target-from-inventory
Drop in a target. Drop in your freezer. We pick the path.
Six workflows are scored in parallel — Gateway (BP/LR and MultiSite), Gibson Assembly, Golden Gate (fragments and gRNA cloning), restriction cloning, SDM, and synthesis. We rank by success_estimate / (1 + work_estimate), hand off to the right designer with pre-resolved arguments, and validate the in-silico product against your target to confirm it preserves function.
workflowsuccesswork
Gateway0.940.25PICK
Gibson0.880.450.61
Golden Gate0.860.400.61
sgRNA cloning0.820.550.53
SDM0.740.620.46
Synthesis1.000.920.52