An arXiv preprint introduces InstructPro2, a method for de novo protein design that combines natural language instructions with biological grounding. The abstract frames the work around a central goal of computational biology: generating proteins with desired functions.

According to the abstract, natural language can express diverse design objectives, but existing protein generation methods often support only narrow objectives. InstructPro2 appears to target that gap, though the provided abstract is cut off before describing the method's specifics or results.

Because only the opening of the abstract is available, this digest cannot report on benchmarks, architecture, or comparisons. The paper is listed as a new announcement on arXiv under identifier 2610.19592.