Shift Bioscience Unveils Improved Virtual Cell Model Ranking to Accelerate Gene Target Discovery

Shift Bioscience, a biotechnology company at the forefront of cell rejuvenation research, has announced the release of a new study that proposes a significantly improved approach to ranking virtual cell models used in gene discovery. The findings promise to accelerate the company’s therapeutic pipeline aimed at combating age-related diseases by enhancing the accuracy and reliability of virtual cell modelling.

The study, led by Lucas Paulo de Lima Camillo, Head of Machine Learning at Shift Bioscience, introduces novel metrics and calibration techniques designed to better assess the performance of virtual cell models trained on single-cell RNA sequencing (scRNA-seq) data. These models are critical tools for simulating how cells respond to gene perturbations, offering a virtual alternative to laborious and time-intensive wet lab experiments.

Despite the promise of virtual cell models in high-throughput gene screening, past benchmarking efforts using standard performance metrics have revealed a surprising limitation: even top models often underperform compared to the simple average of all cells in a dataset. Shift’s new study attributes this issue to misleading signals from weak perturbations and control biases in experimental data.

To overcome these challenges, the team at Shift developed a suite of enhancements that allow for more meaningful evaluations. These include:

• DEG-weighted scoring to emphasize biologically significant gene changes,

• Positive and negative baseline calibrations for clearer performance comparisons, and

• DEG-aware optimization objectives to focus model training on relevant cellular shifts.

Together, these refinements enable researchers to better identify models that truly capture the biological effects of gene perturbations, paving the way for faster and more accurate target discovery.

The company believes this innovation marks a critical step toward the future of drug discovery in aging-related diseases. With improved virtual models, Shift Bioscience can streamline the identification of gene targets with rejuvenation potential, significantly reducing the time and cost associated with bringing new therapies to the clinic.

This latest advancement builds on Shift Bioscience’s growing momentum in the field of cellular rejuvenation. Earlier this year, BioFocus covered the company’s landmark discovery of a breakthrough single-gene target capable of driving safe cellular rejuvenation, a major milestone in age-related therapeutic research. To learn more about this, read the full article here.