Source 1primary paper2025PPR
Toolkit/ML Int&in
ML Int&in
Taxonomy: Technique Branch / Method. Workflows sit above the mechanism and technique branches rather than replacing them.
Summary
ML Int&in is a machine learning–guided computational design method for identifying unnatural split sites in the fast split inteins gp41-1 and NrdJ-1. In the cited preprint, these designs yielded functional split intein variants with reduced fragment affinity and supported blue-light-activatable protein splicing systems that controlled Cre recombinase in mammalian cells.
Usefulness & Problems
Why this is useful
This method is useful for engineering conditional split inteins whose activity can be externally regulated while retaining productive protein splicing. In the cited study, it enabled blue light–dependent control of Cre recombinase and was further applied to spatially control apoptosis through localized expression of truncated BID and caspase-8.
Source:
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Source:
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
Problem solved
The method addresses the challenge of finding noncanonical split sites in fast split inteins that remain functional but have reduced spontaneous fragment association. This property supported construction of light-activatable intein systems for regulated protein splicing in mammalian cells.
Source:
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Source:
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
Published Workflows
Objective: Engineer conditional split inteins for precise spatiotemporal protein regulation in mammalian cells using machine learning–guided split-site design and light-controlled fragment proximity.
Why it works: The workflow is based on first generating split intein variants with reduced mutual affinity, then restoring productive association conditionally by using a light-inducible heterodimerization system to control fragment proximity.
protein trans-splicing by split inteinslight-induced physical proximity of split intein fragmentsmachine learning–guided split-site predictionlight-inducible heterodimerization
Stages
- 1.ML-guided split-site prediction(in_silico_filter)
This stage identifies redesigned split configurations expected to weaken spontaneous fragment association while retaining function, creating a basis for conditional control.
Selection: Predicted unnatural split sites in gp41-1 and NrdJ-1 that would generate functional variants with reduced mutual fragment affinity.
- 2.Conditionalization by light-controlled proximity(functional_characterization)
This stage tests whether reduced-affinity split inteins can be switched on by externally controlling fragment proximity with light.
Selection: Engineered split intein variants were coupled to a light-inducible heterodimerization system to create conditional inteins.
- 3.Mammalian-cell functional validation with Cre(confirmatory_validation)
This stage confirms that the engineered light-activatable inteins can control a functional protein output in mammalian cells.
Selection: Blue light–dependent control of Cre recombinase activity in mammalian cells.
- 4.Downstream application to spatial apoptosis control(confirmatory_validation)
This stage demonstrates that the conditional intein system can drive a spatially resolved cellular phenotype downstream of Cre control.
Selection: Use of the blue-light-controlled system to spatially control apoptosis via localized expression of tBID and caspase-8.
Steps
- 1.Predict unnatural split sites in gp41-1 and NrdJ-1 using ML Int&indesign algorithm
Generate candidate split intein variants with reduced mutual fragment affinity.
Prediction is performed first to choose redesigned split configurations before building conditional light-responsive systems.
- 2.Create conditional inteins by coupling reduced-affinity fragments to a light-inducible heterodimerization systemengineered conditional split inteins
Make intein activity dependent on externally controlled fragment proximity.
This follows split-site redesign because reduced mutual affinity creates the opportunity to gate association through an inducible proximity module.
- 3.Test blue light–dependent control of Cre recombinase activity in mammalian cellsengineered switches under test
Validate that the engineered inteins can control a functional protein output in mammalian cells.
A functional cellular output is tested after conditional inteins are built to confirm that light-controlled splicing translates into regulated protein activity.
- 4.Exploit the system to spatially control apoptosis via localized expression of tBID and caspase-8upstream control modules enabling downstream phenotype
Demonstrate a spatially resolved downstream cellular application of the light-controlled intein system.
This application-level test follows Cre control validation to show that the system can drive a meaningful spatial cellular phenotype.
Taxonomy & Function
Primary hierarchy
Technique Branch
Method: A concrete computational method used to design, rank, or analyze an engineered system.
Mechanisms
blue light-dependent activationreduction of mutual fragment affinitysplit intein-mediated protein splicingTechniques
Computational DesignTarget processes
No target processes tagged yet.
Implementation Constraints
The method was used to design split variants of gp41-1 and NrdJ-1 for mammalian-cell protein splicing applications. The supplied evidence indicates integration into blue-light-activatable systems controlling Cre recombinase, but it does not provide construct architecture, photoreceptor components, illumination parameters, or delivery details.
The evidence provided here comes from a single 2025 preprint and summarizes applications rather than detailed benchmarking of model performance, design success rate, or failure modes. Quantitative comparisons, independent replication, and broader validation across additional inteins, cargos, or cell types are not described in the supplied evidence.
Validation
Cell-freeBacteriaMammalianMouseHumanTherapeuticIndep. Replication
Supporting Sources
Ranked Claims
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
Light-activatable gp41-1 and NrdJ-1 inteins enabled blue light–dependent control of Cre recombinase activity in mammalian cells.
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
The blue-light-controlled intein-Cre system was exploited to spatially control apoptosis via localized expression of truncated BID and caspase-8.
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
Reduced mutual affinity in engineered split intein fragments was harnessed to create conditional inteins by controlling fragment proximity with a light-inducible heterodimerization system.
Reduced mutual affinity in engineered split intein fragments was harnessed to create conditional inteins by controlling fragment proximity with a light-inducible heterodimerization system.
Reduced mutual affinity in engineered split intein fragments was harnessed to create conditional inteins by controlling fragment proximity with a light-inducible heterodimerization system.
Reduced mutual affinity in engineered split intein fragments was harnessed to create conditional inteins by controlling fragment proximity with a light-inducible heterodimerization system.
Reduced mutual affinity in engineered split intein fragments was harnessed to create conditional inteins by controlling fragment proximity with a light-inducible heterodimerization system.
Reduced mutual affinity in engineered split intein fragments was harnessed to create conditional inteins by controlling fragment proximity with a light-inducible heterodimerization system.
Reduced mutual affinity in engineered split intein fragments was harnessed to create conditional inteins by controlling fragment proximity with a light-inducible heterodimerization system.
Reduced mutual affinity in engineered split intein fragments was harnessed to create conditional inteins by controlling fragment proximity with a light-inducible heterodimerization system.
Approval Evidence
1 source1 linked approval claimfirst-pass slug ml-int-in
Using the ML Int&in algorithm, we predicted unnatural split sites in two of the fastest and most efficient split inteins, gp41-1 and NrdJ-1, to generate functional variants with fragments of reduced mutual affinity.
Source:
engineering resultsupports
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
Source:
Comparisons
Source-backed strengths
The reported designs produced functional split variants of both gp41-1 and NrdJ-1, indicating applicability across more than one intein scaffold. The resulting systems enabled blue light–dependent regulation of Cre activity in mammalian cells and supported spatial control of apoptosis in a localized illumination context.
Source:
ML Int&in predicted unnatural split sites in gp41-1 and NrdJ-1 that generated functional split intein variants with reduced mutual fragment affinity.
Ranked Citations
- 1.