Cross-scale modeling of bioreactor cultures using Raman spectroscopy
Aspect of the disclosure relate to methods of assessing a bioreactor culture that involve determining a culture parameter of the manufacturing-scale bioreactor culture using a model that relates a Raman spectrum to the culture parameter. Related bioreactor system are also provided.
1. A method of assessing a bioreactor culture, the method comprising:
(i) obtaining a Raman spectrum of a manufacturing-scale bioreactor culture using a probe configured inline with the bioreactor; and
(ii) determining a culture parameter of the manufacturing-scale bioreactor culture using a multivariate partial least squares (PLS) model that relates the Raman spectrum to the culture parameter, wherein the model was developed based on one or more test bioreactor cultures;
wherein the working volume of the manufacturing-scale bioreactor culture is in a range of 1,000 L to 100,000 L;
wherein the working volumes of the one or more test bioreactor cultures are in a range of 0.1 L to 100 L;
wherein the model was built from spectral data collected during fed-batch cell culture processing from the one or more test bioreactor cultures by exporting the data from one or more respective Raman probes; and
wherein each of the one or more Raman probes is configured for obtaining a Raman spectrum of the respective bioreactor culture during cell growth.
2. The method of claim 1 , wherein
(i) the test bioreactor culture is in a range of 1 L to 5 L; and/or
(ii) the test bioreactor culture is is in a range of 50 L to 100 L.
3. The method of claim 1 , wherein the culture parameter is a level of glucose, glutamate, ammonia or lactate in the culture.
4. The method of claim 3 , wherein
(a) the culture parameter is a level of glucose, and wherein the model has:
(i) a root mean square error of estimation in a range of 0.50 g/L to 1 g/L, and/or
(ii) a root mean square error of cross validation in a range of 0.50 g/L to 1 g/L, and/or
(iii) a root mean square error of prediction in a range of 0.50 g/L to 1.5 g/L, and/or
(iv) an average percentage error of up to 10%;
(b) the culture parameter is a level of lactate, and wherein the model has:
(i) a root mean square error of estimation in a range of 0.10 g/L to 0.20 g/L, and/or
(ii) a root mean square error of cross validation in a range of 0.10 g/L to 0.20 g/L, and/or
(iii) a root mean square error of prediction in a range of 0.10 g/L to 0.20 g/L, and/or
(iv) an average percentage error of up to 20%;
(c) the culture parameter is a level of glutamate, and wherein the model has;
(i) a root mean square error of estimation in a range of 0.10 mM to 0.20 mM, and/or
(ii) a root mean square error of cross validation in a range of 0.10 mM to 0.40 mM, and/or
(iii) a root mean square error of prediction in a range of 0.40 mM to 1.5 mM, and/or
(iv) an average percentage error of up to 35%; or
(d) the culture parameter is a level of ammonium, and wherein the model has:
(i) a root mean square error of estimation in a range of 0.20 mM to 0.40 mM, and/or
(ii) a root mean square error of cross validation in a range of 0.20 mM to 0.50 mM, and/or
(iii) a root mean square error of prediction in a range of 0.40 mM to 1.5 mM, and/or
(iv) an average percentage error of up to 20%.
5. The method of claim 1 , wherein the culture parameter is the osmolality of the culture.
6. The method of claim 5 , wherein the model has:
(i) a root mean square error of estimation in a range of 5 mOsm/kg to 15 mOsm/kg, and/or
(ii) a root mean square error of cross validation in a range of 10 mOsm/kg to 15 mOsm/kg, and/or
(iii) a root mean square error of prediction in a range of 10 mOsm/kg to 25 mOsm/kg, and/or
(iv) an average percentage error of up to 10%.
7. The method of claim 1 , wherein the volume of the manufacturing-scale bioreactor culture is in a range of 2000 L to 20000 L.
8. The method of claim 1 , wherein the volume of the manufacturing-scale bioreactor culture is in a range of 2000 L to 15000 L.
9. A method comprising:
(i) obtaining a Raman spectrum of a first bioreactor culture of a first volume using a probe configured inline with the bioreactor; and
(ii) determining a culture parameter of the first bioreactor culture using a multivariate partial least squares (PLS) model, which model was developed based on a second bioreactor culture of a second volume, that relates the Raman spectrum to the culture parameter, wherein the second volume is in a range of 0.1% to 10% of the first volume;
wherein the model was built from spectral data collected during fed-batch cell culture processing from the one or more test bioreactor cultures by exporting the data from one or more respective Raman probes; and
wherein each of the one or more Raman probes is configured for obtaining a Raman spectrum of the respective bioreactor culture during cell growth.
10. The method of claim 1 , wherein the model was further developed based on at least one bioreactor culture of substantially the same volume as the manufacturing-scale bioreactor culture.
11. The method of claim 1 , wherein
(i) the Raman spectrum comprises spectral signal in the 200cm- 1 to 3400cm- 1 wavenumber range;
(ii) the Raman spectrum comprises spectral signal in the visible, near infrared, infrared, near ultraviolet, or ultraviolet (UV) range;
(iii) the Raman spectrum is obtained using Surface Enhanced Raman Spectroscopy (SERS), resonance Raman spectroscopy, tip-enhanced Raman spectroscopy, polarized Raman spectroscopy, stimulated Raman spectroscopy, transmission Raman spectroscopy, spatially offset Raman spectroscopy, difference Raman spectroscopy, Fourier Transform (FT) Raman spectroscopy, or hyper Raman spectroscopy; and/or
(iv) the Raman spectrum is obtained using a Raman analyzer configured with a laser or other suitable light source configured to operate at wavelengths in a range of 325 nm to 1064 nm.
12. A bioreactor system comprising:
a bioreactor chamber configured for containing a manufacturing-scale bioreactor culture;
a probe configured for obtaining a Raman spectrum of the manufacturing-scale bioreactor culture; and
a computer configured for determining a culture parameter of the manufacturing-scale bioreactor culture, wherein the computer comprises:
an input interface configured to receive information indicative of the Raman spectrum obtained from the probe;
at least one processor programmed to evaluate a model that relates the Raman spectrum to the culture parameter, wherein the model was developed based on one or more test bioreactor cultures, and
an output interface configured to output a signal indicative of the determined culture parameter;
wherein the working volume of the manufacturing-scale bioreactor culture is in a range of 1,000 L to 100,000 L;
wherein the working volumes of the one or more test bioreactor cultures are in a range of 0.1 L to 100 L;
wherein the model was built from spectral data collected during fed-batch cell culture processing from the one or more test bioreactor cultures by exporting the data from one or more respective Raman probes; and
wherein each of the one or more Raman probes is configured for obtaining a Raman spectrum of the respective bioreactor culture during cell growth.
13. The bioreactor system of claim 12 , wherein the output comprises a feedback control signal for controlling operation of a device for altering the culture parameter or wherein the device for altering the culture parameter is a pump or valve configured to control flow, into or out from the bioreactor culture, of a medium comprising one or more culture components.
14. The bioreactor system of claim 12 , wherein
(i) the test bioreactor culture is in a range of 1 L to 5 L; and/or
(ii) the test bioreactor culture is in a range of 50 L to 100 L; and/or
(v) the model is a partial least squares model.
15. The bioreactor system claim 12 , wherein the culture parameter is a level of glucose, glutamate, ammonia or lactate in the culture, or wherein the culture parameter is the osmolality of the culture.
16. The bioreactor system of claim 12 , wherein
(a) the culture parameter is a level of glucose, and wherein the model has:
(i) a root mean square error of estimation in a range of 0.50 g/L to 1 g/L, and/or
(ii) a root mean square error of cross validation in a range of 0.50 g/L to 1 g/L, and/or
(iii) a root mean square error of prediction in a range of 0.50 g/L to 1.5 g/L, and/or
(iv) an average percentage error of up to 10%;
(b) the culture parameter is a level of lactate, and wherein the model has:
(i) a root mean square error of estimation in a range of 0.10 g/L to 0.20 g/L, and/or
(ii) a root mean square error of cross validation in a range of 0.10 g/L to 0.20 g/L, and/or
(iii) a root mean square error of prediction in a range of 0.10 g/L to 0.20 g/L, and/or
(iv) an average percentage error of up to 20%;
(c) the culture parameter is a level of glutamate, and wherein the model has
(i) a root mean square error of estimation in a range of 0.10 mM to 0.20 mM, and/or
(ii) a root mean square error of cross validation in a range of 0.10 mM to 0.40 mM, and/or
(iii) a root mean square error of prediction in a range of 0.40 mM to 1.5 mM, and/or
(iv) an average percentage error of up to 35%; and/or
(d) the culture parameter is a level of ammonium, and wherein the model has:
(i) a root mean square error of estimation in a range of 0.20 mM to 0.40 mM, and/or
(ii) a root mean square error of cross validation in a range of 0.20 mM to 0.50 mM, and/or
(iii) a root mean square error of prediction in a range of 0.40 mM to 1.5 mM, and/or
(iv) an average percentage error of up to 20%.
17. The bioreactor system of claim 12 , wherein the culture parameter is the osmolality of the culture and the model has:
(i) a root mean square error of estimation in a range of 5 mOsm/kg to 15 mOsm/kg, and/or
(ii) a root mean square error of cross validation in a range of 10 mOsm/kg to 15 mOsm/kg, and/or
(iii) a root mean square error of prediction in a range of 10 mOsm/kg to 25 mOsm/kg, and/or
(iv) an average percentage error of up to 10%.
18. The bioreactor system of claim 12 , wherein
(i) the Raman spectrum comprises spectral signal in the 200cm- 1 to 3400cm-1 wavenumber range;
(ii) the Raman spectrum comprises spectral signal in the visible, near infrared, infrared, near ultraviolet, or ultraviolet (UV) range;
(iii) the Raman spectrum is obtained using Surface Enhanced Raman Spectroscopy (SERS), resonance Raman spectroscopy, tip-enhanced Raman spectroscopy, polarized Raman spectroscopy, stimulated Raman spectroscopy, transmission Raman spectroscopy, spatially offset Raman spectroscopy, difference Raman spectroscopy, Fourier Transform (FT) Raman spectroscopy, or hyper Raman spectroscopy; and/or
(iv) the probe comprises a Raman analyzer configured with a laser or other suitable light source configured to operate at wavelengths in a range of 325 nm to 1064 nm.
19. The bioreactor system of claim 12 , wherein the model was further developed based on at least one bioreactor culture of substantially the same volume as the manufacturing-scale bioreactor culture.
20. The method of claim 12 , wherein the volume of the manufacturing-scale bioreactor culture is in a range of 2000 L to 20000 L.
21. The method of claim 12 , wherein the volume of the manufacturing-scale bioreactor culture is in a range of 2000 L to 15000 L.