Robust near-infra-red spectroscopic probe for dynamic monitoring of critical nutrient ratio in microbial fermentation processes
Near infra-red spectroscopy (NIRS) measurements for bioprocess monitoring and control, are integral to process analytical technology (PAT) initiatives by EMEA and US-FDA. Yet, NIRS is not widely practiced in challenging microbial fermentation processes. We present a practical approach to develop NIRS models for linoleic acid (LA), oleic acid (OA) and ammonia which are critical nutrients in lipstatin fermentation by Streptomyces toxitricini. The lipstatin productivity was enhanced and steadied by dynamic monitoring and control of critical nutrient ratio (CNR) of LA to ammonia. The NIRS models were used to develop a novel, soft probe for CNR as an alternative to laborious, hourly, off-line analyses. The calibration was designed with typical data for four industrially useful microbes. The approach enabled direct use of spectra for a generally applicable model with distinct wave number optima of 6250–5555 cm−1 (LA), 6666–5882 cm−1 (OA), 6800–6300 cm−1 (ammonia). The standard errors of calibration and prediction were 1.5 × 10−3 g L−1, 1.6 × 10−3 g L−1, 1.1 ppm, and 8.9 × 10−4 g L−1, 1.8 × 10−2 g L−1, 3.6 ppm, respectively, for the respective nutrients. The robustness of probe is evident from the low mean percentage error of 2.3% for prediction of CNR at low concentration ranges of 0.02–0.24 g L−1 and 0.21–0.56 g g−1 for LA and CNR, respectively.
► Optimal critical nutrient ratio (CNR) improves lipstatin yield from Streptomyces toxitricini. ► A near infra-red spectroscopy (NIRS) soft-probe for CNR was developed. ► High accuracy at atypically low concentrations indicates robustness of the probe. ► CNR probe is an alternative to expensive, arduous and hourly conventional analysis. ► The CNR probe was developed to improve bioprocess monitoring and control.
Journal: Biochemical Engineering Journal - Volume 71, 15 February 2013, Pages 47–56