Optimization of processes by multivariate datanalysis

DAu – Dansk Automationsselskab

October 2nd 2014

Henrik Toft, Process Analytics

Hillerød, Denmark

Overview of presentation

2

Biogen Idec, Hillerød, Denmark

► Introduction

► Traditional biologics manufacturing

► Future biologics manufacturing

► Advanced Process Control at

Biogen Idec

► The multivariate advantage and

how we use it

► Chromatography review example

► Automatic process control

► How to get to the next level

► Q & A

Traditional Biologics Manufactring

3

►Lab and pilot scale experiments to identify process

parameters setpoint

►Large scale process verification runs to verify process

parameters and product quality

►Filing of production process (parameters)

►Regular process trend reporting to verify that process is

in a state of control

Product release based on end-point testing

Future Biologics Manufacturing

4

►Lab and pilot scale experiments conducted using Design

of Experiments to establish Design Space†

►Large scale process verification runs to verify design

space and product quality

►Filing of production process (design space)

►Continous process verification to verify that process is in

a state of control

Product released real time

† Applying uncertainty estimation e.g. through Monte-Carlo simulation

Advanced Process Control at Biogen Idec

5

Data-information pyramid

SBOL, Discoverant,

Delta V,

Trending Program Management, Availability, Analysis, Interpretation

SPC, trends,

Shewhart, Offline MVA

Feedback, Alarms,

Trending controls

SBOL models, MVA,

Transition, Cpk, Golden Batch

MPC, SBOL back to Delta V,

Virtual Sensors, PAT tools

BASIC Monitoring

BASIC Control

ADVANCED Monitoring

ADVANCED Control

Real time

Quality

DATA

INFORMATION

Pro

ce

ss

Un

de

rsta

nd

ing

Highest level of

Understanding/

knowledge –

LEANEST State

Centralized Data Management System

6

Batch Record

Bioreactor Purification. Drug

Substance Inoculation Raw Material

Data/Information

Flow

Material Flow

OPM LIMS

Interface

System allows user to

access data from any

system

Built-in analytics provide

for control charts, CpK

analysis, etc.

PI

Data Management System (Discoverant)

User

Drug

Product

Delta V

Advanced Process Control at Biogen Idec

7

►Establishment of infrastructure for data centralization and real time process monitoring:

►2003: 1st Discoverant Hierarchy & 1st Cell Culture SBOL †

►2006: SBOL on Manufacturing Floor

►2007: 1st Purification SBOL model

►2013: Approved patent on “Systems and Methods for Evaluating Chromatography Column Performance”, US008410928B2, 02Apr2013

►Built a culture of advanced process monitoring

† SIMCA-Batch On-Line

The Multivariate Advantage

8

The two variables are correlated The information is found NOT in the individual signals, but in the correlation pattern through MVA A lot of outliers are not detected unless all the variables are analysed together

Advanced Monitoring through SBOL

9

Multivariate Plot

Contribution Plot

Univariate Plot

Multivariate Plot

SBOL – easy monitoring everywhere

10

► In the facility large monitors are present at

several locations:

► In the inocolutionation room

► In the seed train area

► In the bioreactor hall

► In the purificaiton area

► In office areas

► SBOL status included in shift-handover

reports

► Simple physical setup with local PC for

each extended large monitors

► Not all process steps are monitored by

SBOL models

Easy monitoring with SBOL large wall screen

SBOL catches – selected examples

11

Multivariate Chromatogram Analysis

12

►Motivation

► Chromatograms contain a lot of information that we are not utilizing

► Exploit already available continuous data

► UV, conductivity, pH, pressure, volumetric flow

►Goals

► Non-subjective, quantitative method for evaluating chromatograms

►Example: Analysis of elution peak UV tracing

► MVA model using discrete parameters that describe the

characteristics of the elution UV peak

Elution UV Chromatogram Analysis

13

►Residual host cell protein was showing a gradual upward

trend in the drug substance

►Purification step was the suspected root cause

►Investigation was launched:

►Started by building a multivariate process model

(approximately 60 variables covering cell culture and

purification)

Disection of UV elution peak chromatograms

14

14

UV

0

0.1

0.2

0.3

0.4

0.5

0.6

0 200 400 600 800 1000 1200 1400 1600 1800 2000

UV

1

2

3

4

5

1 2 3 4 5

YV

arP

S(H

CP

)

YPredPS[2](HCP)

HCP Model2.M10 (PLS), HCP Model, PS-HCP Model2

YPredPS[Last comp.](HCP)/YVarPS(HCP)

RMSEP = ---

y=0.9141*x+0.1952R2=0.8139

SIMCA-P+ 11.5 - 5/5/2008 3:48:16 PM

Multivariate model to predict HCP*

15

* Host Cell Protein

Q2=0.73

Further correlation analysis indicated: Chromatogram shapes changed as resin aged

• Eight chromatogram parameters adequately predicts HCP level in the Drug Substance

Chrotogram MVA Summary

16

►Chromatogram MVA enabled successful root cause

investigation of the HCP upward shift:

►Multivariate analysis of cell culture and chromatogram

parameters indicated that chromatogram parameters

were strongly correlated with HCP

►Further correlation analysis indicated that resin age

contributed to the chromatogram shape shift

►Lab studies confirmed that resin age contributed to the

column performance shift and the gradual upward trend

of HCP

Automatic Process Control

17

►Controlling nutrient feeding:

► Using multivariate sensors it is possible to monitor ”chemistry”

inside the bioreacotr e.g. glucose, lactose, ammonia and more.

►Realtime monitoring of (viable) cell density:

► Feeding accoring to the number of live cells to enable extented

bioreactor growth phase and hence increase titer per batch.

► Transfer between seed train bioreactors.

►Set up automated systems base on advanced sensors to

automatically run the batch based on recipie and sensor feedback.

How to get to the next level

18

►Upgrade from SIMCA-Batch On-Line (SBOL)

to SIMCA-online

►Implement use of spectroscopic sensors for

real-time process control and monitoring

►Implementation of real-time prediction models

for product quality forecasting

►Work towards full implementation of

continous process verification

Acknowledgement

19

Manufacturing

• Joydeep Ganguly

• Robert Genduso

• Ben Gilbert

• Ed Goodreau

• Andre Walker

• Sarah Yuan

• Lilong Huang

Engineering

• Jeff Simeone

• Jorg Thommes

• Jennifer Mitchell

Technical Development

• Doug Cecchini

• John Pieracci

Thank you for listening – Questions?

20