retentionprediction

The software tool 'retentionprediction' is a solution for improving compound identification rates in various chromatographic conditions typical of metabolomics experiments. This is achieved by using measured gradient retention times of standard solutes spiked into the sample, along with their known isocratic retention vs. eluent composition relationships to determine effective gradient and flow rate profiles by back-calculation.

One of the key challenges in accurately calculating gradient retention times on a wide range of HPLC systems using a single set of isocratic retention data is the need to properly consider the precise shape of both the gradient and flow rate profiles produced by each instrument. This can be a difficult and time-consuming task. Still, the approach used by 'retentionprediction' simplifies the process by utilizing the measured gradient retention times of standard solutes spiked into the sample.

The retention "projections" of 20 other solutes using these back-calculated profiles were highly accurate, with an error rate as low as 0.23% of the gradient time and an R2 up to 0.99996. This level of accuracy provides significant benefits in the identification of compounds, as a quadrupole MS can identify 38-fold more compounds out of a simulated mixture of 7307, and an FTICR-MS can improve its identification rate nearly two-fold with the same mixture.