Estimating developmental states of tumors and normal tissues using a linear time-ordered model

Area of Science:

• Developmental biology
• Cancer research
• Computational biology

Background:

• Tumor cells display aberrant states, with evidence suggesting distinct developmental stages during tumorigenesis.
• Normal development and stem cell differentiation are highly conserved, ordered processes.
• The 'cancer attractor' concept theoretically links normal development and tumorigenesis in a cell state space, but empirical description remains challenging.

Purpose of the Study:

• To develop a novel computational model for describing the relationship between normal development and tumorigenesis.
• To analyze developmental states of tumor cells from a developmental timescale perspective.
• To investigate the potential for tumor cells to revert to earlier developmental stages.

Main Methods:

• Application of a novel time-ordered linear model utilizing a co-bisector to represent vector directions.
• Transformation of time-course gene expression profiles into 'cell developmental state lines'.
• Analysis of developmental states in tumor samples compared to normal tissues.

Main Results:

• The model successfully described developmental trajectories as lines within a cell state space.
• Mouse liver and ovarian tumors exhibited similarities to early developmental stages.
• Human glioma and ovarian tumor cells demonstrated a shift towards a 'younger' developmental state.

Conclusions:

• The time-ordered linear model effectively captures linear developmental trajectories and gene expression changes over time.
• Tumorigenesis involves distinct developmental states that vary across different tissues.
• Cancer cells may exhibit a reversal of developmental progression.