Statistical tests for intra-tumour clonal co-occurrence and exclusivity

Area of Science:

• Oncology
• Computational Biology
• Genetics

Background:

• Tumour progression is an evolutionary process characterized by intra-tumour heterogeneity due to evolving clones.
• Interactions between tumour clones can significantly impact disease progression and treatment outcomes.
• Identifying co-occurring or exclusive mutation pairs suggests synergistic or antagonistic clonal interactions.

Purpose of the Study:

• To develop a novel statistical framework, GeneAccord, for identifying gene pairs altered in distinct tumour subclones.
• To assess the calibration and power of the GeneAccord framework.
• To investigate clonal interactions in acute myeloid leukaemia (AML) using single-cell sequencing data.

Main Methods:

• Development of the GeneAccord statistical testing framework.
• Analysis of framework calibration and power, comparing it to baseline methods.
• Application of GeneAccord to single-cell sequencing data from 123 acute myeloid leukaemia patients.

Main Results:

• GeneAccord effectively controls type I errors by accounting for clonal evolutionary dependencies.
• Analysis of 123 acute myeloid leukaemia patients identified 1 clonally co-occurring gene pair.
• Eight clonally exclusive gene pairs were identified, predominantly involving key signalling pathway genes.

Conclusions:

• GeneAccord is a robust method for detecting gene pairs altered in distinct subclones, essential for understanding tumour evolution.
• The identified gene pairs provide insights into clonal interactions in acute myeloid leukaemia.
• Clonally exclusive gene pairs frequently involve critical signalling pathways, suggesting their importance in tumour subclone dynamics.