TY - JOUR

T1 - Use of a regression model to study host-genomic determinants of phage susceptibility in MRSA

AU - Zschach, Henrike

AU - Larsen, Mette V.

AU - Hasman, Henrik

AU - Westh, Henrik

AU - Nielsen, Morten

AU - Międzybrodzki, Ryszard

AU - Jónczyk-Matysiak, Ewa

AU - Weber-Dąbrowska, Beata

AU - Górski, Andrzej

PY - 2018

Y1 - 2018

N2 - Staphylococcus aureus is a major agent of nosocomial infections. Especially in methicillinresistant strains, conventional treatment options are limited and expensive, which has fueled a growing interest in phage therapy approaches. We have tested the susceptibility of 207 clinical S. aureus strains to 12 (nine monovalent) different therapeutic phage preparations and subsequently employed linear regression models to estimate the influence of individual host gene families on resistance to phages. Specifically, we used a two-step regression model setup with a preselection step based on gene family enrichment. We show that our models are robust and capture the data’s underlying signal by comparing their performance to that of models build on randomized data. In doing so, we have identified 167 gene families that govern phage resistance in our strain set and performed functional analysis on them. This revealed genes of possible prophage or mobile genetic element origin, along with genes involved in restriction-modification and transcription regulators, though the majority were genes of unknown function. This study is a step in the direction of understanding the intricate host-phage relationship in this important pathogen with the outlook to targeted phage therapy applications.

AB - Staphylococcus aureus is a major agent of nosocomial infections. Especially in methicillinresistant strains, conventional treatment options are limited and expensive, which has fueled a growing interest in phage therapy approaches. We have tested the susceptibility of 207 clinical S. aureus strains to 12 (nine monovalent) different therapeutic phage preparations and subsequently employed linear regression models to estimate the influence of individual host gene families on resistance to phages. Specifically, we used a two-step regression model setup with a preselection step based on gene family enrichment. We show that our models are robust and capture the data’s underlying signal by comparing their performance to that of models build on randomized data. In doing so, we have identified 167 gene families that govern phage resistance in our strain set and performed functional analysis on them. This revealed genes of possible prophage or mobile genetic element origin, along with genes involved in restriction-modification and transcription regulators, though the majority were genes of unknown function. This study is a step in the direction of understanding the intricate host-phage relationship in this important pathogen with the outlook to targeted phage therapy applications.

KW - Bacterial phage resistance

KW - MRSA

KW - Phage therapy

KW - Regression modeling

U2 - 10.3390/antibiotics7010009

DO - 10.3390/antibiotics7010009

M3 - Journal article

C2 - 29382143

AN - SCOPUS:85041359959

VL - 7

JO - Antibiotics

JF - Antibiotics

SN - 2079-6382

IS - 1

M1 - 9

ER -