Biology subjects

Mucha, P. J.

Publications and source records attributed to Mucha, P. J..

4 recordsLinked to original sources

bioRxiv2022

A robust core architecture of functional brain networks supports topological resilience and cognitive performance in aging

Aging is associated with gradual changes in cognition, yet some individuals exhibit protection against aging-related cognitive decline. The topological characteristics of brain networks that support protection against cognitive decline in aging are unknown. Here, we investigated whether the robustness of brain networks, queried via the delineation of the brains core network structure, supports superior cognitive performance in healthy aging individuals (n=320, ages 60-90). First, we decomposed each subjects functional brain networks using k-shell decomposition, finding that cognitive function is associated with more robust connectivity of core nodes, primarily within the frontoparietal control network (FPCN). Next, we find that the resilience of core brain network nodes, within the FPCN in particular, relates to cognition. Finally, we show that the degree of segregation in functional networks mediates relationships between network resilience and cognition. Together, these findings suggest that brain networks balance between robust core connectivity and segregation to facilitate high cognitive performance in aging.

neuroscience↗

bioRxiv2021

Differential Compositional Variation FeatureSelection: A Machine Learning Framework with LogRatios for Compositional Metagenomic Data

The demand for tight integration of compositional data analysis and machine learning methodologies for predictive modeling in high-dimensional settings has increased dramatically with the increasing availability of metagenomics data. We develop the differential compositional variation machine learning framework (DiCoVarML) with robust multi-level log ratio bio-marker discovery for metagenomic datasets. Our framework makes use of the full set of pairwise log ratios, scoring ratios according to their variation between classes and then selecting out a small subset of log ratios to accurately predict classes. Importantly, DiCoVarML supports a targeted feature selection mode enabling researchers to define the number of predictors used to develop models. We demonstrate the performance of our framework for binary classification tasks using both synthetic and real datasets. Selecting from all pairwise log ratios within the DiCoVarML framework provides greater flexibility that can in demonstrated cases lead to higher accuracy and enhanced biological insight.

bioinformatics↗

bioRxiv2021

Pathogenic potential in catheter associated Escherichia coli is associated with separable biofilm and virulence gene determinants

Urinary catheterization facilitates Escherichia coli colonization of the urinary tract and increases infection risk. While specific pathotypes are well-recognized for some E. coli infections, it is unclear whether strain-specific characteristics among E. coli are associated with infection risk in catheterized patients. Here we used comparative genomics and a simulated catheter biofilm model to compare strains associated with catheter-associated urinary tract infection (CAUTI) and catheter-associated asymptomatic bacteriuria (CAASB). CAUTI was associated with a phylotype B2 sub-clade dominated by the multidrug resistant ST131 lineage, while CAASB isolates were genetically more diverse. Catheter-associated biofilm formation was widespread but quantitatively variable among isolates. Network community analysis resolved distinct groups of genes associated with infection or biofilm formation, with iron acquisition-associated genes prominent throughout. Using a reporter construct and targeted mutagenesis, we detected a biofilm phenotype for the ferric citrate transport (Fec) system, the most prominent correlate of high catheter biofilm formation in these patients. In mixed cultures, catheter biofilms formed by some CAASB strains suppressed catheter colonization by ST131 CAUTI isolates. These results are consistent with a paradigm in which catheter biofilm-associated genes increase infection risk in strains with a high pathogenic potential and decrease infection risk through niche exclusion in strains with low pathogenic potential.

microbiology↗

bioRxiv2020

FastPG: Fast clustering of millions of single cells

Current single-cell experiments can produce datasets with millions of cells. Unsupervised clustering can be used to identify cell populations in single-cell analysis but often leads to interminable computation time at this scale. This problem has previously been mitigated by subsampling cells, which greatly reduces accuracy. We built on the graph-based algorithm PhenoGraph and developed FastPG which has the same cell assignment accuracy but is on average 27x faster in our tests. FastPG also has higher cell assignment accuracy than two other fast clustering methods, FlowSOM and PARC. AvailabilityFastPG is available here: https://github.com/sararselitsky/FastPG

bioinformatics↗