Optimization in Biological Systems: The Legacy of Engineering Complex Processes
Mohammad Mazharul Islam
Author
03/30/2021
Added
112
Plays
Description
Optimization is an engineering methodology that uses a mathematical formulation of a design problem to support selection of the optimal design among many alternatives subject to realistic constraints. Engineers have been optimizing mechanical, chemical, physiochemical, and computational processes since the beginning of time. Starting from the invention of calculus to the design of the most complex industrial machines of our time, we have been optimizing every aspect of the design, operation and control of such processes. The history of engineering has thus been one with the development and implementation of numerous classical optimization techniques.
We have continued the legacy of classical optimization in designing biological processes that leverages the plasticity of living systems inherited through evolution. It enables biotechnologists to steer metabolism to many different directions ranging from strain development for chemicals and materials production, drug targeting in pathogens, prediction of enzyme functions, pan-reactome analysis, modeling interactions among multiple cells or organisms, and understanding human diseases. A growing number of computational strain design procedures relying upon mathematical optimization frameworks have emerged benefiting from the rapid advancements in the reconstruction of genome‐scale metabolic models, thus addressing the challenge of identifying and quantifying genetic/environmental interventions and minimizing the counteractions of the organisms in response to them. These large-scale models together with constraint-based optimization methods represent a key foundational advance in Systems Biology and Metabolic Engineering and are crucial for sustainable development in food, pharmaceuticals and bioproduction of the future.
Searchable Transcript
Toggle between list and paragraph view.
- [00:00:00.560]My name is Mohammad.
- [00:00:01.670]I'm working with Professor Rajib Saha
- [00:00:02.677]in the Chemical and Biomolecular Engineering department.
- [00:00:06.420]In this poster,
- [00:00:07.470]I'm highlighting the legacy of engineering optimization
- [00:00:10.540]that is carried on in modeling and engineering
- [00:00:13.270]of biological systems for human benefit.
- [00:00:17.230]In every field of engineering,
- [00:00:18.810]optimization is the central component
- [00:00:20.590]of any sort of decision making.
- [00:00:22.480]Be it scheduling an efficient supply chain,
- [00:00:25.170]finding the shortest route,
- [00:00:26.750]optimal design of load carrying structures
- [00:00:29.210]or improving the efficiency of electricity transmission
- [00:00:31.900]and distribution systems.
- [00:00:34.070]As an engineer or decision maker,
- [00:00:35.930]in any complex process
- [00:00:37.170]we often perform different sorts of optimization algorithms.
- [00:00:40.980]The process takes as input,
- [00:00:43.230]a mathematically explicit model of the system
- [00:00:46.060]and then optimizes for a decision variable
- [00:00:48.460]subject to realistic parameters and constraints,
- [00:00:51.340]which then goes through experimental verification.
- [00:00:55.090]Now, similar to in engineering,
- [00:00:57.530]the living organisms do their own optimization
- [00:01:00.940]because they have to survive
- [00:01:02.330]in their environmental conditions that they live in.
- [00:01:05.270]Here, I'm showing a pipeline network analogy
- [00:01:07.920]that resembles the metabolism in a living cell.
- [00:01:11.270]The flow of fluid is analogous
- [00:01:13.380]to the distribution of the input carbon sources
- [00:01:15.670]like glucose or carbon dioxide.
- [00:01:18.000]While the pipes resemble
- [00:01:19.333]that the actions and the joints are metabolites.
- [00:01:23.500]The nutrients our cell takes as an input,
- [00:01:26.210]goes through this reaction network
- [00:01:27.940]and at the end produces the necessary components
- [00:01:30.780]for growth called biomass.
- [00:01:32.740]In addition, the cell sometimes produces some byproducts
- [00:01:36.400]which can be different than the product that we desire.
- [00:01:40.480]We can redesign or engineer this pipe network
- [00:01:43.200]to produce our desired product
- [00:01:45.250]and minimize the by-product
- [00:01:46.780]while keeping the growth sustained.
- [00:01:49.060]In a pipeline, we achieve this by shutting off
- [00:01:52.150]or opening up valves in different sections
- [00:01:54.780]to redirect the flow of the fluid.
- [00:01:57.050]Similarly, we can use regulations
- [00:01:59.280]that up regulates or down regulates genes in a cell
- [00:02:02.690]that shifts metabolism in the direction that we want.
- [00:02:06.300]What genes we want to modify or engineer
- [00:02:09.210]needs to be identified first
- [00:02:11.380]through optimization of a model cell
- [00:02:13.260]which we call the metabolic model of an organism.
- [00:02:17.650]This is a toy example of a model cell
- [00:02:19.830]where we can see the cellular boundary
- [00:02:22.130]as well as the internal and external reactions.
- [00:02:25.830]One assumption in metabolic modeling
- [00:02:27.280]is that a pseudo steady-state assumption
- [00:02:30.140]where we assume that the metabolism is happening
- [00:02:32.850]in a very fast rate or pseudo steady-state.
- [00:02:36.200]That means we can write the mass balance equations
- [00:02:38.520]for the metabolite in our model
- [00:02:40.340]and end up with a system of linear equations.
- [00:02:43.520]We can then use different optimization algorithms
- [00:02:45.970]to simulate the metabolism of the cellular model.
- [00:02:48.950]One of the most popular one is,
- [00:02:50.610]Flux Balance Analysis or FBA.
- [00:02:53.390]It tries to maximize the growth of an organism
- [00:02:55.640]subject to this mass balance constraints,
- [00:02:58.180]minimum and maximum bounce on reaction rates
- [00:03:01.060]based on thermodynamics
- [00:03:02.790]and environmental constraints.
- [00:03:05.450]In the next few slides,
- [00:03:06.710]I'll show you some examples from my research
- [00:03:09.450]where metabolic modeling and optimization is used
- [00:03:12.070]in biological systems.
- [00:03:14.840]First, a human pathogen, staphylococcus aureus.
- [00:03:18.570]Staphylococcus is a highly antibiotic resistant organism
- [00:03:22.050]and therefore we need novel therapeutics
- [00:03:24.050]to combat this pathogen.
- [00:03:27.010]As part of that mission,
- [00:03:28.270]I recently developed a new genome scale metabolic model
- [00:03:31.520]of staphylococcus aureus that we call iSA863.
- [00:03:35.740]To do that,
- [00:03:36.573]I assembled the initial reaction set
- [00:03:38.510]from the previously published models,
- [00:03:40.880]checked the different biochemical databases
- [00:03:42.960]and included the missing gene annotations
- [00:03:45.690]while formed balance shape on each of the reactions
- [00:03:48.780]and ensured that the model is consistent
- [00:03:50.970]in terms of thermodynamics
- [00:03:52.580]and energy production capabilities.
- [00:03:55.090]In addition, I used a tool called GrowMatch
- [00:03:57.550]to reconcile the experimentally observed
- [00:03:59.580]and model predicted gene essentiality.
- [00:04:02.240]And finally, I used condition specific
- [00:04:04.710]and mutation specific regulation information
- [00:04:07.300]to further define the model.
- [00:04:10.010]I am now employing a multi-level optimization framework
- [00:04:12.790]to identify a minimal number of reactions
- [00:04:15.630]which must be regulated to emulate the phenotype
- [00:04:18.210]that staphylococcus can adapt to,
- [00:04:20.570]given environmental and genetic perturbations.
- [00:04:23.820]This will give us important effect limiting
- [00:04:26.570]and regulatory points
- [00:04:28.120]that can be targeted for therapeutic purposes.
- [00:04:31.760]I'm also looking into synthetic lethal targets
- [00:04:34.270]which are combination of gene knockouts
- [00:04:36.640]that can restrict growth of these pathogen
- [00:04:39.300]and therefore can be used for developing new drugs.
- [00:04:43.840]This is another project
- [00:04:45.070]where I used the concept of optimization
- [00:04:47.220]in understanding heat stress response in rice.
- [00:04:51.000]As we all know, rice is the most consumed food crop
- [00:04:53.840]feeding half of the world's population.
- [00:04:56.040]But elevated temperatures during
- [00:04:57.680]early seed developmental stage
- [00:04:59.430]can adversely the affect,
- [00:05:00.690]the seed size, quality, nutritional content and yield.
- [00:05:04.350]So we attempted to answer the question of,
- [00:05:07.710]what drives these responses in rice seed?
- [00:05:10.640]And if there are a handful
- [00:05:11.960]of global regulatory genes
- [00:05:13.810]that are responsible for the genome-wide changes?
- [00:05:17.610]To do that, we subjected rice plants
- [00:05:20.360]to moderate heat stress of 35 degrees Celsius
- [00:05:23.460]and collected RNA samples from both controlled
- [00:05:26.060]and stressed plants at different time points.
- [00:05:29.830]I then identified the heat-stress responsive genes
- [00:05:32.580]to differential expression
- [00:05:34.330]and did a Pearson correlation analysis
- [00:05:36.540]to find which genes are co-expressed
- [00:05:39.200]with the greatest number of genes in the network.
- [00:05:41.810]These are shown as the orange dots.
- [00:05:44.100]These are our potential candidates for global regulators.
- [00:05:48.200]And then I used my new optimization algorithm
- [00:05:51.320]called MiReN
- [00:05:52.320]that inferred the most influential global regulatory genes
- [00:05:55.950]controlling the expression of other heat responsive genes.
- [00:05:59.930]So, in summary, I tried to convince you
- [00:06:03.430]that we can and do leverage the optimization
- [00:06:06.590]that is inherent in biological systems.
- [00:06:09.850]And for that we utilize genome-scale metabolic models.
- [00:06:13.460]These models can drive
- [00:06:14.880]new therapeutic drug target identification,
- [00:06:17.260]as well as understanding regulatory mechanisms at play
- [00:06:20.770]during abiotic stresses.
- [00:06:23.670]These projects were performed
- [00:06:25.030]in the Systems and Synthetic Biology Laboratory
- [00:06:27.470]led by Professor Saha.
- [00:06:29.450]I also collaborated with
- [00:06:30.610]Professor Vinai Thomas from UNMC,
- [00:06:33.115]Professor Walia from Agronomy and Horticulture
- [00:06:35.580]and Professor Fernando from Animal Science department.
- [00:06:39.410]Thank you for listening.
- [00:06:40.580]If you have any question, I'll be happy to answer.
The screen size you are trying to search captions on is too small!
You can always jump over to MediaHub and check it out there.
Log in to post comments
Embed
Copy the following code into your page
HTML
<div style="padding-top: 56.25%; overflow: hidden; position:relative; -webkit-box-flex: 1; flex-grow: 1;"> <iframe style="bottom: 0; left: 0; position: absolute; right: 0; top: 0; border: 0; height: 100%; width: 100%;" src="https://mediahub.unl.edu/media/16106?format=iframe&autoplay=0" title="Video Player: Optimization in Biological Systems: The Legacy of Engineering Complex Processes" allowfullscreen ></iframe> </div>
Comments
0 Comments