The purpose of the conducted project is to develop bioinformatics methods for detecting transmission pathways for pathogens that can cause infections in intensive care units, and undersanding transmission mechanisms of pathogenic species by applying these methods to real-life data obtained from intensive care units. An innovative bioinformatics method that can be used to investigate hospital infection transmission mechanisms, which is an important health problem, is proposed. The project takes an interdisciplinary approach, combining a number of fields that have experienced important advances and inventions in the last few years: genome biotechnology, metagenomics and microbiome science, computational epidemiology, digital signal processing and machine learning.
The study first conducted controlled experiments to develop bioinformatics methods by deliberately causing microbial transmission among objects in intensive care unit environment. Microbial swab samples from the objects were sequenced using next-generation sequencing, and algorithms predicting “among which microbiomes and in what mechanistics are the transmissions occuring” based on the constructed microbiome profiles. Firstly, digital signal processing methods such as matrix completition and Robust Principal Component Analysis were considered to recognize the contamination content within microbiomes. Consequently, regularized regression algorithms were developed to estimate the transmission network. After identifying and optimizing the most successful algorithm versions tested on the training data, they were run on the real-life nosocomial data. Swab samples gathered from Erciyes University Hospital Internal Medicine ICU wiere analyzed using the computational source tracking method and potential infection transmission mechanisms are mapped. Moreover, metagenomics analysis of the predicted environmental pathogens were performed and phylogeny, antibiotics resistance profiles and computational sorce tracking were studied.
ICU Sequencing data can be found on:
With the following NCBI accession IDs:
BioProject ID: PRJNA561526
The code and software packages are available at:
This project is funded by TUBİTAK (The Scientific and Technological Research Council of Turkey) 3501 Program (Project no: 115E998)