Machine learning algorithm to improve estimation and integration of single-cell data

Assistant Professor Daifeng Wang, with research group members Noah Cohen Kalafut and Xiang Huang, published a paper titled “Joint variational autoencoders for multimodal imputation and embedding” in Nature Machine Intelligence on May 29.

The algorithm, called JAMIE for Joint Variational Autoencoders for Multimodal Imputation and Embedding, incorporates cross-modal imputation to estimate values or missing data and integrates different modalities of a cell. This integration of multimodal data allows investigators to fully understand cellular function.

JAMIE, which is open source, addresses two data challenges: making sense of large, complex data sets in an integrative way, and estimate larger quantities of missing data.

See the Waisman Center press release for more information.

Congratulations to Daifeng and his team!