Related Experiment Video
Updated: Mar 24, 2026
Accurate and Phenol Free DNA Sexing of Day 30 Porcine Embryos by PCR
Published on: February 14, 2016
Deep learning-based early screening for porcine embryos with different developmental potential
Yongjiang Yang1, Haoxing Li2, Dengfeng Bi3
1State Key Laboratory of Organ Regeneration and Reconstruction, Institute of Zoology, Chinese Academy of Sciences, Beijing, 100101, China; Institute for Stem Cell and Regeneration, Chinese Academy of Sciences, Beijing, 100101, China; University of Chinese Academy of Sciences, Beijing, 100049, China; Beijing Institute for Stem Cell and Regenerative Medicine, Beijing, 100101, China.
Researchers developed a deep learning model, MaxViT_T, to predict early-stage porcine embryo development for improved artificial reproduction technology (ART) efficiency. This AI tool enhances embryo screening by overcoming challenges posed by high lipid content in in vitro embryos.
Area of Science:
- Veterinary Medicine
- Animal Reproduction
- Bioinformatics
Background:
- Porcine artificial reproduction technology (ART) efficiency is limited by the difficulty of predicting in vitro embryo developmental potential.
- High cytoplasmic lipid content in embryos obscures optical clarity, hindering microscopic assessment and reliable prediction.
Purpose of the Study:
- To develop and evaluate deep learning models for predicting the developmental potential of early-stage porcine embryos.
- To establish a novel embryonic droplet culture system for collecting high-quality image data.
Main Methods:
- Collected 10,041 bright-field images of porcine parthenogenetically activated (PA) embryos (1-cell to blastocyst stage).
- Developed and benchmarked multiple deep learning models, including MaxViT_T, using curated developmental outcome annotations.
- Validated model performance across different experimental batches with varying blastocyst formation rates.
Main Results:
- Identified MaxViT_T as the most efficient deep learning model for predicting porcine embryo development.
- MaxViT_T achieved peak prediction performance at the 4-cell stage.
- The model demonstrated consistent prediction efficiency across diverse experimental conditions.
Conclusions:
- The MaxViT_T model accurately predicts the developmental potential of early-stage porcine embryos.
- This AI-driven approach offers a novel method for screening high-quality embryos in ART.
- Improved embryo selection can enhance the overall efficiency of porcine reproductive technologies.
