MSCA: a spectral comparison algorithm between time series to identify protein-protein interactions.

Area of Science:

• Computational Biology
• Bioinformatics
• Biophysics

Background:

• Pathogen-host interactions are key to understanding infections and developing therapeutics.
• Identifying targeted host proteins is crucial but challenging and expensive.
• Existing methods for understanding virulence mechanisms are time-consuming and costly.

Purpose of the Study:

• To develop a novel statistical method for identifying functional relationships between proteins.
• To create an algorithm that compares protein time series based on physicochemical properties.
• To predict protein-protein interactions (PPIs) and pathogen-host interactions (PHIs).

Main Methods:

• Developed the Multiple Spectral Comparison Algorithm (MSCA), inspired by BLASTP.
• Implemented MSCA in R code for statistical analysis of protein data.
• Utilized hypothesis testing on spectral densities of physicochemical property time series.

Main Results:

• MSCA demonstrated high accuracy (70%) in detecting PPIs and PHIs with a 0.7 threshold.
• The algorithm successfully identified known interactions of human proteins (MAGI1, SCRIB, JAK1) and virulence proteins (ROP16, ROP18, ROP17, ROP5).
• Simulation studies confirmed MSCA's effectiveness with unequal time series.

Conclusions:

• The MSCA algorithm effectively identifies protein similarities and interactions using spectral density comparisons.
• This method offers a cost-effective and efficient approach to studying protein relationships.
• Potential new interactions were proposed, including human GBP and CREB as substrates for ROP protein complexes.