Genetic Variability Impacts Genotoxic and Transcriptome Responses in the Human Colon after the Consumption of Processed Red Meat Products and Those with Added Phytochemical Extracts.

The PHYTOME study investigated the effect of consuming processed meat products on outcomes related to colorectal cancer risk without testing the impact of genetic variability on these responses. This research aims to elucidate the genetic impact on apparent total N-nitroso compound (ATNC) excretion, colonic DNA adduct formation, ex vivo-induced DNA damage, and gene expression changes in colon biopsies of healthy participants. Through a systematic literature review, candidate polymorphisms were selected and then detected using TaqMan and PCR analysis. The effect of genotype on study outcomes was determined via a linear mixed model and analysis of variance. Machine learning was used to evaluate relative allele importance concerning genotoxic responses, which established a ranking of the most protective alleles and a combination of genotypes (gene scores). Participants were grouped by GSTM1 genotype and differentially expressed genes (DEGs), and overrepresented biological pathways were compared between groups. Stratifying participants by ten relevant genes revealed significant variations in outcome responses. After consumption of processed red meat, variations in NQO1 and COMT impacted responses in ATNC levels (µmol/L) (+9.56 for wildtype vs. heterozygous) and DNA adduct levels (pg/µg DNA) (+1.26 for variant vs. wildtype and +0.43 for variant vs. heterozygous), respectively. After phytochemicals were added to the meat, GSTM1 variation impacted changes in DNA adduct levels (-6.12 for deletion vs. wildtype). The gene scores correlated with these responses and DEGs were identified by GSTM1 genotype. The altered pathways specific to the GSTM1 wildtype group included 'metabolism', 'cell cycle', 'vitamin D receptor', and 'metabolism of water-soluble vitamins and co-factors'. Genotype impacted both the potential genotoxicity of processed red meat and the efficacy of protective phytochemical extracts.

Low mitochondrial DNA copy number in buffy coat DNA of primary open-angle glaucoma patients.

Primary open-angle glaucoma (POAG) is characterized by optic nerve degeneration and irreversible loss of retinal ganglion cells (RGCs). The pathophysiology is not fully understood. Since RGCs have a high energy demand, suboptimal mitochondrial function may put the survival of these neurons at risk. In the present study, we explored whether mtDNA copy number or mtDNA deletions could reveal a mitochondrial component in POAG pathophysiology. Buffy coat DNA was isolated from EDTA blood of age- and sex-matched study groups, namely POAG patients with high intraocular pressure (IOP) at diagnosis (high tension glaucoma: HTG; n = 97), normal tension glaucoma patients (NTG, n = 37), ocular hypertensive controls (n = 9), and cataract controls (without glaucoma; n = 32), all without remarkable comorbidities. The number of mtDNA copies was assessed through qPCR quantification of the mitochondrial D-loop and nuclear B2M gene. Presence of the common 4977 base pair mtDNA deletion was assessed by a highly sensitive breakpoint PCR. Analysis showed that HTG patients had a lower number of mtDNA copies per nuclear DNA than NTG patients (p-value <0.01, Dunn test) and controls (p-value <0.001, Dunn test). The common 4977 base pair mtDNA deletion was not detected in any of the participants. A lower mtDNA copy number in blood of HTG patients suggests a role for a genetically defined, deficient mtDNA replication in the pathology of HTG. This may cause a low number of mtDNA copies in RGCs, which together with aging and high IOP, may lead to mitochondrial dysfunction, and contribute to glaucoma pathology.

Predicting missing proteomics values using machine learning: Filling the gap using transcriptomics and other biological features.

Proteins are often considered the main biological element in charge of the different functions and structures of a cell. However, proteomics, the global study of all expressed proteins, often performed by mass spectrometry, is limited by its stochastic sampling and can only quantify a limited amount of protein per sample. Transcriptomics, which allows an exhaustive analysis of all expressed transcripts, is often used as a surrogate. However, the transcript level does not present a high level of correlation with the corresponding protein level, notably due to the existence of several post-transcriptional regulatory mechanisms. In this publication, we hypothesize that the missing protein values in proteomics could be predicted using machine learning regression methods, trained with many features extracted from transcriptomics, including known translational regulatory elements such as microRNAs and circular RNAs. After considering different machine learning algorithms applied on two different splitting strategies, we report that random forest can predict proteins in new samples out of transcriptomics data with good accuracy. The proposed pre-processing and model building scripts can be accessed on GitHub: https://github.com/jochotecoa/ml_proteomics.