Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

In Food research international (Ottawa, Ont.)

This work reports on the metabolic fingerprinting of ten new races of Capsicum annuum cv. jalapeño using 1H NMR based metabolomics coupled to machine learning projections. Ten races were classified and evaluated according to their differential metabolites, variables of commercial interest and by multivariate data analysis/machine learning algorithm. According to our results, experimental races of jalapeño peppers exhibited differences in carbohydrate, amino acid, nucleotide and organic acid contents. Forty-eight metabolites were identified by 1D and 2D NMR and the differential metabolites were quantified by qNMR. Principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) separated the studied races into two groups. The group A included the races Colosus, Emperador, Fundador and Rayo whereas the group B included the races Don Benito, SMJ 1416, SMJ 1417, SMJ 1423, SMJ 145 and STAM J0904. OPLS-DA revealed that levels of citric acid in group A were higher than in group B, while the levels of asparagine, fumaric acid, GABA, glucose, malic acid, pyruvic, quinic acid, sucrose and tryptophan were higher in the group B. Remarkably, ascorbic acid was exclusively found in the race Colosus. Random forest model revealed the diversity of the experimental races and the similarity rate with the well-established races. The most relevant variables used to generate a model were length, weight, yield, width, xylose content and organic acids content.

Ramírez-Meraz Moisés, Méndez-Aguilar Reinaldo, Hidalgo-Martínez Diego, Villa-Ruano Nemesio, Zepeda-Vallejo L Gerardo, Vallejo-Contreras Fernando, Hernández-Guerrero Claudia J, Becerra-Martínez Elvia

2020-Dec

Jalapeno pepper, Machine learning, Multivariate statistical analysis, NMR, Random forest