BindCraft2 (BC2) is available on Tamarind Bio on day one for commercial and academic use. The new workflow substantially expands on the original BindCraft, with support for 10 binder modalities, multiple targets and target states, and more control over developability properties.
The promise of “pick an epitope, design a binder against it” is becoming more sophisticated. BC2 gives researchers more ways to specify the molecule they want to create, the interactions they want it to make, and the constraints that matter for their program.
BC2 supports VHH, scFv, Fab, and seven other binder modalities. This broadens the design space beyond miniproteins and brings more antibody formats into the same workflow. Researchers can start with a format relevant to their intended application and explore designs within that context. For teams working on antibody discovery, the addition of these formats is a particularly meaningful expansion.
Multi-target support is another major addition. The same binder can be designed against multiple targets, including versions of a protein across species. For example, a campaign could aim for a binder that recognizes both the human target and its counterpart in a species used for preclinical research. This makes the desired binding profile something researchers can specify at the start of design.
BC2 also supports detargeting: incorporating proteins that a binder should avoid into the design objective. Together, targeting and detargeting let researchers express a more complete selectivity goal. A campaign might seek binding across a chosen set of related targets while discouraging interactions with others. These capabilities create a more direct way to design toward the interactions a program needs, with experimental testing still determining the resulting binding profile.
Support for multiple target states adds another dimension. Researchers can account for different conformations of a target within a design campaign. Alongside induced-fit support and the ability to target intrinsically disordered regions (IDRs), this expands the range of target contexts that can be explored.
The workflow also provides controls for prioritizing properties such as humanization and termini placement. These controls bring considerations about the eventual molecule into the design process earlier. Researchers can tune the objectives around their program’s requirements and explore how those preferences affect the candidates generated.
Taken together, these additions expand both the binder formats and the biological questions accessible through the workflow. The design brief can now include the desired modality, multiple intended targets, unwanted interactions, target states, and properties relevant to downstream development. That gives scientists a richer starting point for designing and testing new binders.