CytoGPS: A large-scale karyotype analysis of CML data

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Karyotyping is crucial for diagnosing genetic disorders, including cancers, but relies on manual analysis of visually recorded chromosomal abnormalities.
• Karyotypes are documented using the International System for Human Cytogenetic Nomenclature (ISCN), which is not computer-readable, limiting large-scale data analysis.
• The inability to process ISCN data computationally hinders the full realization of insights from extensive genomic datasets.

Purpose of the Study:

• To develop a computational platform, CytoGPS, for analyzing large volumes of cytogenetic data.
• To convert human-readable ISCN karyotypes into a machine-readable binary format for automated analysis.
• To demonstrate the platform's utility in identifying novel patterns within cancer karyotype data.

Main Methods:

• Developed CytoGPS, a platform utilizing a Loss-Gain-Fusion model to transform ISCN text into a binary format.
• Applied CytoGPS to over 69,000 cancer karyotypes from the Mitelman Database.
• Employed the Jaccard coefficient to compare karyotypes represented as binary vectors for similarity analysis.

Main Results:

• Successfully converted human-readable ISCN karyotypes into a machine-readable binary format.
• Identified novel patterns in 4,968 Mitelman leukemia karyotypes, including the co-occurrence of trisomy 19 and 21.
• Demonstrated CytoGPS's capability for large-scale, comparative analysis of cytogenetic data.

Conclusions:

• CytoGPS effectively transforms cytogenetic data, unlocking potential for advanced genomic research.
• The platform facilitates the discovery of new correlations and patterns in cancer chromosomal aberrations.
• CytoGPS provides a freely accessible tool for researchers to analyze large-scale cytogenetic datasets.