Premium
Denaturing high‐performance liquid chromatography and principal component analysis for identification of DNA point mutations in breast cancer and lymphoma samples
Author(s) -
PerfectoAvalos Yocanxóchitl,
CuevasDíaz Durán Raquel,
Villela Luis,
GarciaGonzalez Alejandro,
DíazDomínguez Ricardo Javier,
Loyo Tania,
GutiérrezMonreal Miguel Ángel,
EsparzaTreviño Juan Manuel,
RochaInclán Carlos,
Rojo Rocío,
CárdenasCantú Eduardo,
PantaléonGarcía Jezreel,
Scott SeanPatrick
Publication year - 2018
Publication title -
journal of chemometrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.47
H-Index - 92
eISSN - 1099-128X
pISSN - 0886-9383
DOI - 10.1002/cem.3053
Subject(s) - principal component analysis , outlier , pattern recognition (psychology) , chromatography , sequence analysis , multivariate statistics , artificial intelligence , computational biology , mathematics , biology , dna , computer science , chemistry , genetics , statistics
DNA mutations are identified by techniques that use the knowledge of the wild‐type DNA sequence and its mutated variant. The involved analytic methods must be accurate, rapid, and sustainable, if a clinical application is pursued. High‐performance liquid chromatography under denaturing conditions is a useful technique to screen mutations. Denaturing high‐performance liquid chromatography resultant chromatograms are suitable for feature extraction analysis with multivariate methods such as principal component analysis. In this work, principal component analysis was applied to analyze the chromatograms from 3 different genes. Fragments with verified wild‐type sequence were used as reference and samples with sequence unknown were tested. A statistical characterization based on Tukey's boxplot equation of principal component scores allowed us to analyze the distance distribution between reference and sample clusters to establish a classification criterion: an outlier could represent a mutated sample, and a typical value could be a wild‐type sample. Identified outliers were further analyzed by sequencing and proved to carry a mutation. From 72 datasets with a total of 4258 injections, we successfully assessed the classification criterion, identifying mutated samples in lymphoma and breast cancer patients with ratio of prediction G mean = [0.89, 1.00]. Compared with sequencing analysis, this procedure reduced time and costs.