Surface protein imputation from single cell transcriptomes by deep neural networks.

Area of Science:

• Computational Biology
• Genomics
• Proteomics

Background:

• Single-cell RNA sequencing (scRNA-seq) provides insights into cellular heterogeneity.
• Cell-surface proteins are crucial for cell function and are key therapeutic targets.
• Directly measuring surface proteins at single-cell resolution is challenging.

Purpose of the Study:

• To develop a computational framework for predicting cell surface protein abundance.
• To leverage scRNA-seq data for inferring protein expression profiles.
• To enable the identification of potential therapeutic targets using transcriptomic data.

Main Methods:

• Proposed a transfer learning framework named cTP-net (cell Transcriptome to Protein prediction).
• Utilized deep neural networks for predicting protein abundances.
• Trained the model on existing single-cell multi-omic datasets.

Main Results:

• Successfully imputed surface protein abundances from scRNA-seq data.
• Demonstrated the capability of cTP-net to predict protein expression profiles.
• Provided a method to bridge the gap between transcriptomic and proteomic data.

Conclusions:

• cTP-net offers a powerful approach to infer cell surface protein expression from scRNA-seq.
• This method can significantly aid in understanding cell biology and discovering therapeutic targets.
• Facilitates the use of widely available scRNA-seq data for protein-level insights.