Microbial strain optimisation for the overproduction of a desired phenotype has been a popular topic in recent years. Gene knockout is a genetic engineering technique that can modify the metabolism of microbial cells to obtain desirable phenotypes. Optimisation algorithms have been developed to identify the effects of gene knockout. However, the complexities of metabolic networks have made the process of identifying the effects of genetic modification on desirable phenotypes challenging. Furthermore, a vast number of reactions in cellular metabolism often lead to a combinatorial problem in obtaining optimal gene knockout. The computational time increases exponentially as the size of the problem increases. This work reports an extension of Bees Hill Flux Balance Analysis (BHFBA) to identify optimal gene knockouts to maximise the production yield of desired phenotypes while sustaining the growth rate. This proposed method functions by integrating OptKnock into BHFBA for validating the results automatically. The results show that the extension of BHFBA is suitable, reliable, and applicable in predicting gene knockout. Through several experiments conducted on
The rapid development of genetic manipulation techniques has made the alteration of microorganisms for different purposes popular in recent years. Genetic manipulation of microorganisms aims to increase the yields of biocompounds or decrease the production of by-products [
Computational algorithms have been developed to identify the gene knockout to obtain improved phenotypes. Burgard et al. developed the first rational modelling framework (known as OptKnock) for introducing a gene knockout, leading to the overproduction of a desired metabolite [
OptKnock uses mixed integer linear programming (MILP) to formulate a bilevel linear optimisation that is a promising method of finding the global optimal solution. OptGene is an extended approach of OptKnock, which formulates the in silico design problem using a Genetic Algorithm (GA) [
A hybrid of BA and FBA (BAFBA) was proposed by Choon et al. [
This paper is organised as follows. First, the materials and experimental setup are described. Then, the problem formulation is introduced, and the details of the BAFBA and the extension of BHFBA are described. Next, experimental results are presented. Then, the obtained results are discussed, reviewing the contributions of this work. Finally, this paper is summarised by providing the main conclusion and addresses future developments.
In this study, we used
We compared the results with those of previous reports in the literature [
The problem of identifying optimal gene knockout from biological models can be formulated as follows. Suppose that a model that contains the stoichiometric matrix
Flux vector:
The chemical elements, ionic charge, and biochemical moieties must be balanced in the stoichiometric matrix. The objective is to find the optimal gene knockout to improve the product yields of industrially important chemicals while sustaining the growth rate of the microorganism. This is commonly performed using linear programming, defined as follows:
Figure
BAFBA flowchart.
Before attempting to propose the extension of BHFBA, it is crucial to find the limitations of the BAFBA [
In this paper, we propose the extension of BHFBA to identify optimal gene knockout. It is proposed to overcome the limitations of BAFBA and previous reports. The extension of BHFBA in our work differs from the BAFBA in local search activities and in validating the results. The extension of BHFBA improves the algorithm by hybridising Hill Climbing algorithm with BAFBA and by integrating OptKnock into BHFBA. Figure
BHFBA flowchart.
Bee representation of metabolic genotype.
We used biomass-product coupled yield (BPCY) as the fitness score in this work. According to Soons et al., metabolic networks can function in living cells under various biological objectives depending on the relevant organism and its genetic and environmental context. However, biological objectives are only applicable for analysing a number of organisms in terms of microbial metabolic engineering. It is desirable to couple the formation of the desired product to growth [
The flow of calculating the fitness function is shown in Figure
The flow for calculating fitness function.
The flow of OptKnock validation.
In this paper, we propose an improved method, extension of BHFBA, to test the performance of the BHFBA. For evaluation, we conduct a benchmarking analysis. However, benchmark functions can only be tested on BH and BA because FBA is an objective function. Hence, we test the benchmark functions on BH and BA in this study. Because BA is used to look for a maximum, the functions are inverted before the algorithm is applied. The De Jong, Martin and Gaddy, Schwefel, and Griewangk functions are used in this study. These functions are a set of common parametric test problems. The simplest test function is De Jong. It is continuous, convex, and unimodal. Martin and Gaddy function is a unimodal function. The Schwefel function is complex, with many local minima. Lastly, Griewangk function has many widespread local minima. However, the location of the minima is regularly distributed. We carry out 100 individual runs to test BH and BA.
Table
Mathematical representation of De Jong, Martin and Gaddy, Schwefel, and Griewangk functions.
Name | Mathematical representation |
---|---|
De Jong |
|
|
|
Martin and Gaddy |
|
|
|
Schwefel |
|
|
|
|
|
Obtained fitness value of all benchmark functions.
Function | Mean | STD | ||
---|---|---|---|---|
BA | BH | BA | BH | |
De Jong |
|
|
|
|
Martin and Gaddy |
|
|
|
|
Schwefel |
|
|
|
|
Griewangk |
|
|
|
|
In this paper, the extension of BHFBA is compared with the previous works: BAFBA, SA + FBA, and the conventional OptKnock. Tables
Comparison between different methods for production of Succinic acid by
Method | Growth rate (1/hr) | BPCY | List of knockout genes |
---|---|---|---|
BHFBA |
|
|
PTAr |
BAFBA [ |
0.62404 | 0.66306 | FUM, PTAr |
SA + FBA [ |
N/A | 0.39850 | ACLD19 |
OptKnock [ |
0.28 | N/A | ACKr, PTAr |
Comparison between different methods for production of Lactic acid by
Method | Growth rate (1/hr) | BPCY | List of knockout genes |
---|---|---|---|
BHFBA |
|
|
FBP, PGK, ACALD |
BAFBA [ |
0.58586 | 3.5656 | GAPD, L_LACD2, PTAr |
SA + FBA [ |
N/A | 0.39850 | ACLD19 |
OptKnock [ |
0.28 | N/A | ACKr, PTAr |
Table
Comparison between different methods for growth rate and BPCY of succinic acid by
Table
Comparison between different methods for production of lactic acid by
We applied the BAFBA to
Table
Comparison between different methods for growth rate and BPCY of ethanol by
Method | Growth rate (1/hr) | BPCY | List of knockout genes |
---|---|---|---|
BHFBA |
|
|
ALAD_L |
BAFBA [ |
122.8861 |
|
ALAD_L |
Comparison between different methods for growth rate and BPCY of ethanol by
Table
Result of implementation of different knockout for production of Ethanol in
Method | Growth rate (1/hr) | BPCY | List of knockout genes |
---|---|---|---|
BHFBA |
|
|
G3PD1 |
BAFBA [ |
9.9313 |
|
MDH, G3PD1 |
Comparison between different methods for production of ethanol by
Table
Comparison between average computational time of BHFBA and BAFBA for 1000 iterations.
Model | Method | Computation time (seconds) |
---|---|---|
|
BHFBA |
|
BAFBA [ |
10253 | |
OptKnock [ |
N/A | |
SA + FBA [ |
N/A | |
|
||
|
BHFBA |
|
BAFBA [ |
22515 | |
OptKnock [ |
N/A | |
SA + FBA [ |
N/A | |
|
||
|
BHFBA |
|
BAFBA [ |
10282 | |
OptKnock [ |
N/A | |
SA + FBA [ |
N/A |
As seen in the results, both the extension of BHFBA and BH performed better than other algorithms. It can be concluded that the ability of the Hill Climbing algorithm to find local optimum improved the performance of the original BA. The original BA has a problem with repetitive iterations of the algorithm in local search, where each bee continues to search until the best possible answer is reached. Our proposed extension of BHFBA solves the problem by implementing Hill Climbing algorithm in the local search and improved the algorithm by integrating OptKnock. The Hill Climbing algorithm is a powerful local search algorithm that attempts to find the best solution by incrementally changing a single element of the solution until no further improvements can be found. The search process is recorded so that the process is not repeated. Furthermore, one of the advantages of the Hill Climbing algorithm is that it can return a valid solution even if it is interrupted at any time before it ends. OptKnock is widely used for
It is crucial to develop more accurate and efficient modelling and optimisation methods in metabolic engineering because they will have a significant impact on commercialised biotechnology engineering, which will lead to substantial economic gains in the production of pharmaceuticals, fuels, and food ingredients. In this paper, the extension of BHFBA is proposed for use in predicting optimal sets of gene deletions to maximise the production of the desired metabolite. The extension of BHFBA improves the performance of the BAFBA by implementing the Hill Climbing algorithm, which is a promising algorithm for finding local optimum. It is extended by integrating OptKnock into BHFBA. Experimental results with
The authors declare that they have no conflict of interests.
The authors would like to thank Malaysian Ministry of Science, Technology and Innovation for supporting this research by an e-science research grant (Grant no. 01-01-06-SF1234). This research is also funded by an Exploratory Research Grant Scheme (Grant no. R.J130000.7807.4L096) and a Fundamental Research Grant Scheme (Grant no. R.J130000.7807.4F190) from Malaysian Ministry of Higher Education.