Integrating patients in time series clinical transcriptomics data.

Area of Science:

• Computational Biology
• Genomics
• Clinical Trials

Background:

• Analyzing time series transcriptomics data from clinical trials presents significant challenges.
• Existing methods often rely on linear, global orderings that fail to capture individual response rates and patient subgroups.

Purpose of the Study:

• To develop a novel method for trajectory inference in large-scale clinical studies using time series transcriptomics data.
• To address the limitations of current methods in handling varying patient response patterns and dynamics.

Main Methods:

• Utilized multi-commodity flow algorithms for trajectory inference.
• Developed a method that integrates data from multiple patients while respecting individual timing restrictions.

Main Results:

• The new method demonstrated improved performance compared to existing approaches on multiple drug datasets.
• Successfully identified novel disease subtypes corresponding to heterogeneous patient response patterns.

Conclusions:

• The developed method offers a more robust approach to analyzing complex clinical trial transcriptomics data.
• This approach enhances the understanding of disease heterogeneity and patient-specific responses.