Soft sensing of intracellular states for CHO cell bioprocessing with Ensemble Kalman Filters

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Asparagine plays a central metabolic role in mammalian cell culture, acting as a nitrogen donor for amino acid and nucleotide biosynthesis, supporting protein translation, and contributing to TCA cycle intermediates that sustain biomass growth and productivity.

“Asparagine plays a central metabolic role in mammalian cell culture, acting as a nitrogen donor for amino acid and nucleotide biosynthesis, supporting protein translation, and contributing to TCA cycle intermediates that sustain biomass growth and productivity [48].”

• cites [48] Kirsch, 2022, Metabolic analysis of the asparagine and glutamine dynamics in an industrial Chinese hamster ovary fed-batch process, Biotechnol. Bioeng. sourcePartly supported

  Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured

May combine two claims.

The EnKF cannot narrow the UDP-Glc band, since UDP-Glc is unobservable and its synthesis is modelled as a fixed function of extracellular glucose that neglects transport and the regulation of glycolytic flux.

“The EnKF cannot narrow this band, since UDP-Glc is unobservable and its synthesis is modelled as a fixed function of extracellular glucose that neglects transport and the regulation of glycolytic flux [52].”

• cites [52] Maldonado, 2014, ATP/ADP ratio, the missed connection between mitochondria and the Warburg effect, Mitochondrion sourceCould not verify

  The Introduction discusses glucose metabolism and the Conclusion discusses regulation of glycolysis by ATP/ADP ratios, but neither addresses the claimed UDP-Glc model. Searches for EnKF, UDP-Glc observability, and UDP-glucose synthesis found no passage supporting the asserted reason the band cannot narrow. (LLM, conf=0.78)

  Experimental — accuracy not yet measured

May combine two claims.

NSDs are synthesised intracellularly from extracellular metabolites such as glucose, glutamine, galactose, and uridine through a sequence of cytosolic enzymatic reactions, and are subsequently consumed in the Golgi apparatus during glycosylation.

“NSDs are synthesised intracellularly from extracellular metabolites such as glucose, glutamine, galactose, and uridine through a sequence of cytosolic enzymatic reactions, and are subsequently consumed in the Golgi apparatus during glycosylation [20,21,25].”

• cites [20] Jimenez del Val, 2011, A dynamic mathematical model for monoclonal antibody N-linked glycosylation and nucleotide sugar donor transport within a maturing Golgi apparatus, Biotechnol. Prog. sourceNot checked — no full text of the source
• cites [21] Jedrzejewski, 2014, Towards controlling the glycoform: A model framework linking extracellular metabolites to antibody glycosylation, Int. J. Mol. Sci. sourcePartly supported

  Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured
• cites [25] Kotidis, 2019, Model-based optimization of antibody galactosylation in CHO cell culture, Biotechnol. Bioeng. sourceNot checked — no full text of the source

May combine two claims.

The mechanistic model systematically overestimated viable cell density for all feeding strategies, even though it was parametrised using data from the same cell line.

“As shown in Fig. 4(a)-(d), the mechanistic model systematically overestimated viable cell density for all feeding strategies, even though it was parametrised using data from the same cell line.”

May combine two claims.

Glucose is corrected in the mean, but its innovations remain far above the consistency target even at the noise cap.

“Glucose is corrected in the mean, but its innovations remain far above the consistency target even at the noise cap, as reported in Supplementary Information C.”

May combine two claims.

UDP-Glc has the widest band of the reported NSDs, because the glucose spread propagates through the submodel.

“As UDP-Glc is synthesised directly from glucose, this spread propagates through the submodel and gives UDP-Glc the widest band of the reported NSDs.”

May combine two claims.

Glutamine carries the narrowest posterior band of the measured states, so its corrected trajectory is the most tightly constrained of the measured states.

“Glutamine also carries the narrowest posterior band of the measured states, so its corrected trajectory is the most tightly constrained of the measured states.”

May combine two claims.

The UDP-GlcNAc uncertainty band remains wide, but the corrected estimate sits well within it and the residual overprediction is much reduced.

“The uncertainty band remains wide, as for the other unobservable NSDs, but the corrected estimate sits well within it and the residual overprediction is much reduced.”

May combine two claims.

The width of the estimated UDP-Gal band scales with the size of the precursor pool: it remains narrow while UDP-Gal is low in early culture and broadens as the pool fills after the later galactose feeds.

“The width of the estimated band scales with the size of the precursor pool.”

• Also stated in: § 3.4.1. Galactosylation precursor dynamics: UDP-Gal · plain

  “It remains narrow while UDP-Gal is low in early culture and broadens as the pool fills after the later galactose feeds, so the reported confidence contracts and expands with the underlying concentration.”

May combine two claims.

Uridine concentrations were quantified using HPLC and therefore exhibited lower measurement noise.

“In contrast, uridine concentrations were quantified using HPLC [9] and therefore exhibited lower measurement noise.”

• cites [9] Del Val, 2013, An optimized method for extraction and quantification of nucleotides and nucleotide sugars from mammalian cells, Anal. Biochem. sourceNot checked — no full text of the source

May combine two claims.

The posterior band includes the measurements throughout the culture, with most data points contained within the ±1σ envelope and a spread-to-error ratio close to unity.

“The posterior band includes the measurements throughout the culture, with most data points contained within the ±1𝜎 envelope and a spread-to-error ratio close to unity.”

May combine two claims.

The open-loop model underpredicts UDP-Glc through the mid-culture phase, mirroring its underprediction of extracellular glucose.

“The open-loop model underpredicts UDP-Glc through the mid-culture phase, mirroring its underprediction of extracellular glucose.”

May combine two claims.

The uncertainty bands are narrowest where the estimate is most confident, as for well-determined states such as glutamine, and substantially wider for structurally mismatched ones such as glucose.

“The uncertainty bands are narrowest where the estimate is most confident, as for well-determined states such as glutamine, and substantially wider for structurally mismatched ones such as glucose.”

May combine two claims.

The EnKF estimate of UDP-Gal is comparable to the open-loop prediction, because the mechanistic model already reproduces the measured trajectory and the assimilated galactose mainly keeps the estimate consistent with it; the filter adds a calibrated uncertainty band rather than a much more accurate mean.

“The EnKF estimate of UDP-Gal is comparable to the open-loop prediction, as shown in Fig. 6(a)-(d), because the mechanistic model already reproduces the measured trajectory and the assimilated galactose mainly keeps the estimate consistent with it.”

• Also stated in: § 3.4.3. Branching precursor with strong inferability: UDP-GlcNAc · plain

  “UDP-Gal is already reproduced well by the open-loop model and gains a calibrated uncertainty band rather than much more accurate mean.”

May combine two claims.

Correcting the glutamine level through assimilation removes most of the UDP-GlcNAc bias, and the EnKF estimate tracks the measurements well across all 4 datasets.

“Correcting the glutamine level through assimilation removes most of this bias, and the EnKF estimate tracks the measurements well across all 4 datasets.”

May combine two claims.

As asparagine is observable, the measurements constrain the estimate and its uncertainty band stays tight, in contrast to the unobservable NSDs.

“As asparagine is observable, the measurements constrain the estimate and its uncertainty band stays tight, in contrast to the unobservable NSDs examined next.”

May combine two claims.

UDP-Gal receives essentially the same additive process noise but is driven by the better-behaved galactose, so its band is correspondingly narrower than that of UDP-Glc.

“UDP-Gal receives essentially the same additive process noise but is driven by the better-behaved galactose, so its band is correspondingly narrower.”

May combine two claims.

The EnKF estimates for galactose aligned closely with experimental measurements while avoiding overfitting of spurious fluctuations, particularly in regions where apparent concentration increases occurred in the absence of feeding events.

“Fig. 4(e)-(h) show the EnKF estimates for galactose aligned closely with experimental measurements while avoiding overfitting of spurious fluctuations, particularly in regions where apparent concentration increases occurred in the absence of feeding events.”

May combine two claims.

The EnKF adjusted glucose trajectories, but glucose retains the widest posterior band of the measured states and the filter reports lower confidence than for other states.

“The EnKF adjusted glucose trajectories accordingly, although glucose retains the widest posterior band of the measured states and the filter reports lower confidence comparing to other states.”

May combine two claims.

The UDP-Glc estimate is the least certain of the reported NSDs, and this uncertainty is inherited from glucose.

“The UDP-Glc estimate is the least certain of the reported NSDs, and this uncertainty is inherited from glucose.”

May combine two claims.

The open-loop model overpredicts UDP-GlcNAc across all feeding strategies, increasingly so towards the end of culture, following its overprediction of glutamine.

“The open-loop model overpredicts UDP-GlcNAc across all feeding strategies, increasingly so towards the end of culture, following its overprediction of glutamine.”

May combine two claims.

The EnKF compensates for the mismatch by assimilating viable cell density measurements, recovering growth trajectories that closely follow the experimental profiles.

“The EnKF effectively compensates for this mismatch by assimilating viable cell density measurements, recovering growth trajectories that closely follow the experimental profiles.”

May combine two claims.

For both galactose and uridine, the ±2σ envelope contains most of the experimental data.

“For both states the ±2𝜎 envelope contains most of the experimental data.”

May combine two claims.

Assay differences are visible in the posterior bands: the galactose band is broad and the filter follows the underlying trend rather than the scattered readings, while the more precise uridine measurements yield a narrower band.

“These assay differences are visible in the posterior bands.The galactose band is broad, and the filter follows the underlying trend rather than the scattered readings.The more precise uridine measurements yield a narrower band.”

May combine two claims.

Measurement assimilation via the EnKF corrected the glutamine deviations and recovered trajectories that closely matched the observed profiles.

“Measurement assimilation via the EnKF corrected these deviations and recovered trajectories that closely matched the observed profiles.”

May combine two claims.

The open-loop model reproduces asparagine early in culture but drifts progressively low thereafter, falling close to zero by the end of every batch, whereas the measured concentration declines only gradually.

“the open-loop model reproduces asparagine early in culture but drifts progressively low thereafter, falling close to zero by the end of every batch, whereas the measured concentration declines only gradually”

May combine two claims.

The NSDs' uncertainty bands are wide, in contrast to the narrow band obtained for observable asparagine.

“Their uncertainty bands are therefore wide, in contrast to the narrow band obtained for observable asparagine.”

May combine two claims.

For the directly measured extracellular states, the EnKF corrected the systematic biases of the open-loop model and attenuated measurement noise, achieving a lower error than the open-loop model for every variable and across all datasets.

“For the directly measured extracellular states, the EnKF corrected the systematic biases of the open-loop model and attenuated measurement noise, achieving a lower error than the open-loop model for every variable and across all datasets.”

May combine two claims.

The system has a strictly unidirectional information structure in which extracellular states drive NSD synthesis, while variations in NSD concentrations do not propagate measurable effects back to the extracellular environment.

“This system structure reflects an inherent biological property characterised by a strictly unidirectional information structure, in which extracellular states drive NSD synthesis, while variations in NSD concentrations do not propagate measurable effects back to the extracellular environment.”

May combine two claims.

The corrections applied to the measured states point to an underlying model bias, which takes a parametric and a structural form.

“The corrections applied to the measured states points to an underlying model bias and takes two forms.”

Classified as probably not a substantial claim.

A simulation-based observability assessment based on the empirical observability Gramian was adopted.

“To address these challenges, a simulation-based observability assessment based on the empirical observability Gramian was adopted.”

May combine two claims.

UDP-GlcNAc is the NSD that benefits most from assimilation and is reconstructed most accurately, because it is coupled to well-tracked glutamine through a well-identified submodel.

“making UDP-GlcNAc the NSD that benefits most from assimilation”

• Also stated in: § 4. Conclusions · plain

  “UDP-GlcNAc, coupled to welltracked glutamine through a well-identified submodel, is reconstructed most accurately.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

UDP-GlcNAc is a branching precursor in N-linked glycosylation and acts as the substrate for several branching and modification reactions.

“UDP-GlcNAc is a branching precursor in N -linked glycosylation and acts as the substrate for several branching and modification reactions [53,54].”

• cites [53] Krambeck, 2005, A mathematical model of N-linked glycosylation, Biotechnol. Bioeng. sourceNot checked — no full text of the source
• cites [54] Kontoravdi, 2018, Computational tools for predicting and controlling the glycosylation of biopharmaceuticals, Curr. Opin. Chem. Eng. sourceNot checked — no full text of the source

May combine two claims.

The confidence of an NSD estimate depends in addition on the uncertainty of its driving state, so an input that is tracked in the mean but structurally mismatched, such as glucose, yields an accurate but weakly constrained NSD, as seen for UDP-Glc.

“The confidence of an estimate depends in addition on the uncertainty of its driving state, so an input that is tracked in the mean but structurally mismatched, such as glucose, yields an accurate but weakly constrained NSD, as seen for UDP-Glc.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The Ensemble Kalman Filter (EnKF) is a Monte Carlo-based extension of the Kalman Filter designed for nonlinear and high-dimensional systems.

“The EnKF is a Monte Carlo-based extension of the Kalman Filter designed for nonlinear and high-dimensional systems [43].”

• cites [43] Evensen, 1994, Sequential data assimilation with a nonlinear quasi-geostrophic model using Monte Carlo methods to forecast error statistics, J. Geophys. Res. sourceNot checked — no full text of the source

May combine two claims.

The ensemble-based formulation ensures that uncertainty propagation remains consistent with the nonlinear process dynamics, allowing variability in observable extracellular metabolite states to be transmitted to intracellular NSD concentrations through mechanistic coupling, without imposing spurious linear sensitivity.

“This ensemble-based formulation ensures that uncertainty propagation remains consistent with the nonlinear process dynamics, allowing variability in observable extracellular metabolite states to be transmitted to intracellular NSD concentrations through mechanistic coupling, without imposing spurious linear sensitivity.”

May combine two claims.

UDP-GlcNAc synthesis kinetics were well identified in the original calibration, since UDP-GlcNAc is chemically stable and comparatively reliable to measure.

“The result is reinforced by the fidelity of the UDP-GlcNAc submodel, since UDP-GlcNAc is chemically stable and comparatively reliable to measure [55], so its synthesis kinetics were well identified in the original calibration.”

• cites [55] Sugiura, 2007, MS analysis of chondroitin polymerization: Effects of Mn2+ ions on the stability of UDP-sugars and chondroitin synthesis, Anal. Biochem. sourceNot checked — no full text of the source

May combine two claims.

The EnKF does not overcome the structural unobservability of NSDs, and the inferred posterior distributions remain dependent on model structure and uncertainty specification.

“The EnKF does not overcome the structural unobservability of NSDs, and the inferred posterior distributions remain dependent on model structure and uncertainty specification.”

May combine two claims.

UDP-Glc is accurate in the mean while weakly constrained, with its confidence limited by the upstream uncertainty in glucose.

“UDP-Glc is therefore accurate in the mean while weakly constrained, with its confidence limited by the upstream uncertainty in glucose.”

• Also stated in: § 3.4.3. Branching precursor with strong inferability: UDP-GlcNAc · plain

  “UDP-Glc is corrected in the mean through the recovered glucose concentration but inherits the uncertainty of the structurally mismatched glucose state, so its estimate is accurate yet weakly constrained.”

May combine two claims.

The progressive underestimation of asparagine by the open-loop model is a downstream signature of the growth bias seen in the measured states, pointing to overestimated growth and asparagine-uptake parameters.

“Asparagine uptake scales with the viable cell density, so this drift is a downstream signature of the growth bias seen in the measured states.”

• Also stated in: § 4. Conclusions · plain

  “The progressive underestimation of asparagine by the open-loop model is a downstream signature of the growth bias, since the overpredicted cell growth overconsumes the pool, and assimilation of the coupled measurements removes it.”

• AssumptionRests on an assumption: Asparagine uptake scales with the viable cell density.

May combine two claims.

The higher-than-predicted glucose accumulation is consistent with the reduced cell growth observed experimentally, compounded by galactose and glucose both being supplied to the culture.

“This behaviour is consistent with the reduced cell growth observed experimentally, compounded by galactose and glucose are both supplied to the culture.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Each scenario was simulated 10 times and the mean trajectory was reported.

“Each scenario was simulated 10 times and the mean trajectory was reported, ensuring repeatability while preserving the stochastic nature of the EnKF.”

Classified as probably not a substantial claim. May combine two claims.

The EnKF represents state uncertainty through an ensemble of full nonlinear system trajectories, and the sample covariance of the ensemble provides a Monte Carlo approximation of the evolving state uncertainty.

“Instead of the Jacobian-based local linearisation and analytical covariance propagation of the EKF, the EnKF represents state uncertainty through an ensemble of full nonlinear system trajectories.The sample covariance of the ensemble provides a Monte Carlo approximation of the evolving state uncertainty.”

May combine two claims.

The overprediction of viable cell density most plausibly reflects an overestimated maximum specific growth rate that propagates to the coupled central-carbon states, since the model was calibrated on lower-passage cultures.

“The model was calibrated on lower-passage cultures, so the overprediction of viable cell density most plausibly reflects an overestimated maximum specific growth rate that propagates to the coupled central-carbon states.”

May combine two claims.

The NSDs are structurally unobservable, a limitation that no estimator can overcome by itself.

“As the observability analysis shows, they are structurally unobservable, a limitation that no estimator can overcome by itself.”

May combine two claims.

Each filter was tuned on a single batch and evaluated on the 3 held-out batches, whose feeding strategies span the galactose and uridine design space.

“Each filter was tuned on a single batch and evaluated on the 3 held-out batches, whose feeding strategies span the galactose and uridine design space [26].”

• cites [26] Kotidis, 2019, Constrained global sensitivity analysis for bioprocess design space identification, Comput. Chem. Eng. sourceNot checked — no full text of the source

May combine two claims.

In the model, asparagine is tightly coupled to the specific growth rate, which in turn governs viable cell density, protein synthesis, and the uptake and secretion of extracellular metabolites.

“In the model, asparagine is tightly coupled to the specific growth rate, which in turn, governs viable cell density, protein synthesis, and the uptake and secretion of extracellular metabolites.”

May combine two claims.

Under galactose supplementation, reduced glucose uptake leads to elevated extracellular glucose concentrations and altered intracellular precursor dynamics.

“Under galactose supplementation, reduced glucose uptake leads to elevated extracellular glucose concentrations and altered intracellular precursor dynamics [47].”

• cites [47] Gramer, 2011, Modulation of antibody galactosylation through feeding of uridine, manganese chloride, and galactose, Biotechnol. Bioeng. sourceNot checked — no full text of the source

May combine two claims.

UDP-Gal is synthesised directly from extracellular galactose and is therefore the NSD most closely coupled to an assimilated measurement.

“UDP-Gal is synthesised directly from extracellular galactose and is therefore the NSD most closely coupled to an assimilated measurement.”

Classified as probably not a substantial claim. The quote only partly states this claim.

The measured-state coefficients of variation were fixed by a filter-consistency criterion requiring the normalised innovation variance to equal unity.

“The measured-state coefficients of variation are fixed by a filterconsistency criterion.At each measurement time the innovation and its predicted covariance are”

Classified as probably not a substantial claim.

The performance of the EnKF-based estimation framework was evaluated using 4 additional fed-batch experiments that were not used during model parametrisation.

“The performance of the EnKF-based estimation framework was evaluated using 4 additional fed-batch experiments that were not used during model parametrisation.”

May combine two claims.

The EKF relies on Jacobian-based linearisation of both the process and measurement models to propagate state covariances and compute correction Kalman gains; when unmeasured states do not influence output measurements, the corresponding sensitivities vanish, leading to degenerate or ill-conditioned covariance updates.

“The EKF relies on Jacobian-based linearisation of both the process and measurement models to propagate state covariances and compute the correction Kalman gains [45].When the unmeasured states do not influence the output measurements, the corresponding sensitivities vanish, leading to degenerate or illconditioned covariance updates.”

• cites [45] Simon, 2006 sourceNot checked — no full text of the source

May combine two claims.

The validation experiments employed the same CHO-T127 cell line, culture medium, and operating conditions as those used for model calibration.

“These experiments were conducted independently but employed the same CHO-T127 cell line, culture medium, and operating conditions as those used for model calibration.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The EnKF represents uncertainty using an ensemble of state trajectories that are evolved forward through the nonlinear process model and updated sequentially as measurements become available.

“Instead of propagating state covariances analytically, the EnKF represents uncertainty using an ensemble of state trajectories that are evolved forward through the nonlinear process model and updated sequentially as measurements become available.”

Probably not a substantial claim: may not pass S3 (load-bearing).

When galactose is supplemented, carbon flux can be redirected towards central energy metabolism, reducing reliance on glucose as the primary carbon and energy source.

“When galactose is supplemented, carbon flux can be redirected towards central energy metabolism, reducing reliance on glucose as the primary carbon and energy source.”

• cites [25] Kotidis, 2019, Model-based optimization of antibody galactosylation in CHO cell culture, Biotechnol. Bioeng. sourceNot checked — no full text of the source
• cites [47] Gramer, 2011, Modulation of antibody galactosylation through feeding of uridine, manganese chloride, and galactose, Biotechnol. Bioeng. sourceNot checked — no full text of the source

May combine two claims.

The EnKF requires neither the sigma-point tuning of the UKF nor the repeated optimisation of moving-horizon estimation, and it scales to higher dimensions more favourably than particle filtering.

“It thereby requires neither the sigma-point tuning of the UKF nor the repeated optimisation of moving-horizon estimation, and it scales to higher dimensions more favourably than particle filtering.”

May combine two claims.

UDP-Glc provides a direct link between central carbon metabolism and glycosylation, primarily governed by extracellular glucose availability.

“UDP-Glc provides a direct link between central carbon metabolism and glycosylation, primarily governed by extracellular glucose availability [51].”

• cites [51] Sha, 2019, An investigation of nucleotide sugar dynamics under the galactose supplementation in CHO cell culture, Process Biochem. sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The measurement-noise covariance was set directly from the data as the experimental-triplicate variance, pooled across the 4 datasets, and was never fitted to the filter residuals.

“The measurement-noise covariance is set directly from the data as the experimental-triplicate variance, pooled across the 4 datasets.This variance reflects measurement precision and is never fitted to the filter residuals.”

Classified as probably not a substantial claim. May combine two claims.

The complete tuning procedure was cross-validated across the 4 fed-batch datasets, with each fold using one dataset as the sole tuning set and applying the calibrated filter to the 3 held-out batches.

“The complete procedure is cross-validated across the 4 fed-batch datasets.In each fold, one dataset serves as the sole tuning set and the calibrated filter is applied to the 3 held-out batches.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Each validation experiment applied a distinct galactose and uridine supplementation profile, selected to span a design space constrained by requirements on product titre and galactosylation quality.

“Each experiment applied a distinct galactose and uridine supplementation profile, selected to span a design space constrained by requirements on product titre and galactosylation quality, as reported in [26].”

• cites [26] Kotidis, 2019, Constrained global sensitivity analysis for bioprocess design space identification, Comput. Chem. Eng. sourceNot checked — no full text of the source

May combine two claims. The quote only partly states this claim.

The unmeasured-state additive variances were calibrated to achieve near-nominal coverage and a spread-to-error ratio of order unity for the states that carry validation measurements.

“The scale factor is selected by calibrating the resulting uncertainty band, so as to achieve near-nominal coverage and a spread-to-error ratio of order unity for the states that carry validation measurements.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Five datasets were used exclusively for parametrisation of the mechanistic model prior to its integration with the EnKF.

“Five datasets were used exclusively for parametrisation of the mechanistic model prior to its integration with the EnKF.”

May combine two claims.

The difference in observability between uridine and galactose arises from model structure: uridine enters the specific death rate and growth inhibition term, while galactose influences dynamics only indirectly through modulation of the specific glucose uptake rate.

“This difference arises from model structure.Uridine enters the specific death rate directly (Eq.( 24), Supplementary Information A) and also the growth inhibition term (Eq.( 23), Supplementary Information A), so its perturbations propagate rapidly through the viable cell density and all growth-coupled states.Galactose, by contrast, influences the dynamics only indirectly through modulation of the”

May combine two claims.

In practice, uncertainty in NSD states accumulates without meaningful correction, while linearisation errors introduced at each update step are not dissipated by measurement feedback, resulting in numerical instability or divergence.

“In practice, uncertainty in NSD states accumulates without meaningful correction, while linearisation errors introduced at each update step are not dissipated by measurement feedback, resulting in numerical instability or divergence.”

Probably not a substantial claim: may not pass S3 (load-bearing).

UDP-Gal provides the mechanistic link between extracellular galactose and uridine supplementation and galactosylation, a key quality attribute of monoclonal antibodies.

“It provides the mechanistic link between extracellular galactose and uridine supplementation and galactosylation, a key quality attribute of monoclonal antibodies [13,50].”

• cites [13] Hodoniczky, 2005, Control of recombinant monoclonal antibody effector functions by Fc N-glycan remodeling in vitro, Biotechnol. Prog. sourceNot checked — no full text of the source
• cites [50] Bheemareddy, 2019, Effect of the total galactose content on complement-dependent cytotoxicity of the therapeutic anti-CD20 IgG1 antibodies under temperature stress conditions, J. Carbohydr. Chem. sourceNot checked — no full text of the source

May combine two claims.

Asparagine is a viable soft-sensing target, and the observability analysis offers a principled basis for selecting such estimable states.

“Asparagine is thus a viable soft-sensing target, and the observability analysis offers a principled basis for selecting such estimable states.”