Development of an Enhanced Total Ion Current Chromatogram Algorithm to Improve Untargeted Peak Detection

Abstract

Accurate analyte peak detection from the background noise is a fundamental step in data analysis. Often, this is initially performed on the total ion current chromatogram (TIC), which is the summed signal from all mass spectral channels. Despite the detection of many of the most abundant peaks within a chromatogram, a large fraction of peaks remains undetected in the standard TIC due to their signal being below the limit of detection. To find peaks obscured by background noise, an untargeted peak detection method termed the "enhanced TIC algorithm" was developed for comprehensive twodimensional gas chromatography coupled with time-of-flight mass spectrometry (GC × GC-TOFMS). The reported algorithm utilizes the entire mass spectral dimension to find regions of analytical signal above a threshold while zeroing the background noise. The resulting chromatographic data is summed together to create the enhanced TIC. The utility of the enhanced TIC algorithm is demonstrated using serial dilutions from a 10 parts-per-thousand (ppth) test mixture. For the chromatograms collected at 1 and 10 parts-per-million (ppm), the enhanced TIC algorithm recovered 62% and 93%, respectively, of the original peaks observed in the 10 ppth mixture, while the standard TIC recovered only 0% and 45%, respectively. The improvement in signal enhancement was also shown on a separation of a yeast cell metabolite extract, where the enhanced TIC found 33−64% more peaks than the standard TIC. Chromatographic simulations with increasing levels of background noise were also conducted to compare the enhanced and standard TICs in the context of statistical overlap theory (SOT). Simulated chromatograms with lower signal-to-noise were more accurately modeled by the SOT after enhanced TIC processing compared to those processed by the standard TIC. The enhanced TIC method demonstrates an immense benefit in peak discovery to improve data analysis efforts.

From Centroided to Profile Mode: Machine Learning for Prediction of Peak Width in HRMS Data

“…However, a recent study has highlighted the potential of the XIC-based feature detection approaches for these techniques. 46 Thus, the presented algorithms, additionally, enable detailed (i.e., XIC based) feature detection in the dataset generated by higher dimension instruments such as comprehensive two-dimensional chromatography coupled with HRMS, which as of today rely on TIC level feature detection due to data complexity. 37 , 47 All the algorithms are developed in an open source and open access manner and consequently are vendor-independent.…”

“…Additionally, one of the outputs of the centroiding algorithm is the mass peak width at half-height, which can be used during data processing workflows (e.g., region of interest detection, XIC-based feature detection algorithms, and setting of mass accuracies for feature identification , ). However, a recent study has highlighted the potential of the XIC-based feature detection approaches for these techniques . Thus, the presented algorithms, additionally, enable detailed (i.e., XIC based) feature detection in the dataset generated by higher dimension instruments such as comprehensive two-dimensional chromatography coupled with HRMS, which as of today rely on TIC level feature detection due to data complexity. , All the algorithms are developed in an open source and open access manner and consequently are vendor-independent.…”

The Often-Overlooked Power of Summary Statistics in Exploratory Data Analysis: Comparison of Pattern Recognition Entropy (PRE) to Other Summary Statistics and Introduction of Divided Spectrum-PRE (DS-PRE)

“…Unfolding changes the data structure from three to two dimensions. , For this image analysis, the 1-norm was applied to the unfolded image, and the value of the 1-norm for each spectrum was assigned to the corresponding pixel. In the case of images, the 1-norm is equivalent to the total signal count in a pixel. , ToF-SIMS image of two different proteins (hemoglobin and bovine serum albumin (BSA)) coated on two different types of fluorescent microspheres. These spheres were deposited onto a clean silicon wafer and air-dried.…”

“…Unfolding changes the data structure from three to two dimensions. , For this image analysis, the 1-norm was applied to the unfolded image, and the value of the 1-norm for each spectrum was assigned to the corresponding pixel. In the case of images, the 1-norm is equivalent to the total signal count in a pixel. , …”

30th Anniversary of comprehensive two‐dimensional gas chromatography: Latest advances

“…An enhanced total ion chromatogram (TIC) algorithm was developed to improve peak detection and identification by identifying regions where the analytical signal is above a given threshold and zeroing the background noise 35 . The improved signal provided by the algorithm was demonstrated on standard test mixtures and on yeast cell metabolite extracts, in which the enhanced TIC found up to 64% more analytes than the classical TIC.…”