Tree reconstruction guarantees from CRISPR-Cas9 lineage tracing data using Neighbor-Joining.

Area of Science:

• Computational Biology
• Genomics
• Evolutionary Biology

Background:

• CRISPR-Cas9 based lineage tracing enables single-cell phylogeny reconstruction from transcriptional data.
• Developing algorithms with theoretical guarantees for tree reconstruction in this context is challenging.

Purpose of the Study:

• To derive a novel tree-reconstruction algorithm with theoretical guarantees for CRISPR-Cas9 lineage tracing.
• To address the realistic scenarios of unknown parameters and missing data in evolutionary models.

Main Methods:

• Utilized Neighbor-Joining (NJ) on moment-matched distances to estimate true tree distances.
• Developed analytical tools to prove theoretical guarantees for the reconstruction algorithm.
• Applied the method to simulated lineage tracing data and real mouse lung cancer data.

Main Results:

• The algorithm achieves theoretical guarantees, aligning with established evolutionary models when parameters are known and data is complete.
• New theory demonstrates reconstruction guarantees are still possible with unknown parameters and missing data.
• Empirical results show improved performance compared to traditional NJ on both simulated and real datasets.

Conclusions:

• The developed algorithm offers a robust approach for phylogenetic tree reconstruction in CRISPR-Cas9 lineage tracing.
• The new theoretical framework extends reconstruction guarantees to realistic scenarios with missing data.
• This work advances computational methods for analyzing complex biological systems and evolutionary processes.