Team:UESTC/nebula

From 2013.igem.org

Revision as of 15:08, 28 October 2013 by BillXue (Talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

hidde original style

Hello,welcome to visit our Wiki, for better view, please use chrome, firefox, safari .etc modern browser.

Nebula

A biological circuit design tool composed of Auto Mode and Manual Mode.

Download


Background

When a man is to begin his own journey of biobricks, he probably needs to turn to the parts registry for information of the parts with which he can construct his own genetically engineered machine. The official site of parts registry is really a terrific work and thousands of biobricks parts have been collected from all around the world. What is more, relatively reasonable classifications of parts have been well characterized and this could be of help for us. Several attempts have already been made in order to simplify the process of designing a biobrick machine, some did well in visualization, but few of them combined the excavated information of classification of parts. Nor could they automatically generate optimized biological circuit according to quality of parts. These indeed reined the efficiency when designing. So the question is raised that a well visualized user - computer interactive and automatic tool that could help us to design the desirable device is in great need.

Nebula (Network of Elaborated Biobricks based on User Locating and Automation) is a biological circuit design tool composed of Auto Mode & Manual Mode. It is freely available for devices powered by IOS. We classified the parts released in 2013 and constructed a database for users to choose what they want. We use AHP (Analytic Hierarchy Process) to score parts and edges (passage linking two parts) according to parts quality including availability, usefulness, sample status, part status and sequencing. According to weight of edges, we present “Index of stability” to users. Users can save the circuits designed in Nebula in case they want to check or change it later.




Algorithm

Interactive Part:

The general idea is to classify the biobricks more intensively in a view of user needs. And with these well-defined and characterized classifications, users can select the classification and sub-classification step by step and at last find the parts they want to get. The definition of the classifications of the parts in this software is mainly based on their biological functions, characteristics as well as the structures and also the existing classification methods on the official registry. After researching the features of the parts, we classified all parts released in 2013. For example, promoters in the parts can be classified into four categories: constitutive, sensitive, association with CDS experience and source. And further, the sub-category sensitive can be classified into the following two sub-categories: by inducer and by repressor, and further the sub-category by inducer can be classified into IPTG, LasI/LasR, RhII/RhIR, metal ion and other inducible promoters. So long as the user knows a little basic backgrounds of his work, he will find all the IPTG-induced promoters in our classification system, which would save a lot of time. Moreover, in this process, our software can connect the parts that the users have chosen and the visualization module will show the structure of the constructs. So in this way, a new biobricks device will be design efficiently. The specific classifications of all the parts released in 2013 are detailed as follows:

Fig.3 The classification of promoters

No classification for RBS.

Fig.4 The classification of CDSs
Fig.5 The classification of terminators
Fig.6 The classification of devices

Automatic Part:

In this part, users only need to determine the inducer and the product. Then, our software will offer users the optimized circuit with the input and output that they designate.

To realize this function, we first collect information of parts released in 2013, including (1) the upstream and downstream parts of each part and device according to the standard of “promoter-RBS-CDS-terminator”; (2) the inducer or repressor of each promoter and device with promoters; (3) the product of each CDS or device with CDSs; (4) the assembly standards each part and device fit; (5) quality of parts and devices including: availability, usefulness, sample status, part status and sequencing.

After collecting all the information, we first use AHP (Analytic Hierarchy Process) to analyze quality of parts and then score every part and device. AHP is a method for organizing and analyzing complex decisions and it is advantageous when factors of the decision are difficult to quantify. The detailed method is as follows:

(1) Construct Model of Hierarchy Structure

This structure includes 3 levels. On the top, it is the level of goal, the score of every part. At the bottom, it is the level of alternatives, the candidate parts. In the medial, it is the level of criteria, the quality of parts. The structure is shown below:

Fig.7 Hierarchy Structure

(2) Construct Pairwise Comparison Matrix

Supposing there are N factors in a certain level, we now need to compare the degree of importance to the upper level of parts with each other and calculate the relative weight of each factor. This comparison is pairwise. Usually we choose 9 scales when comparing these factors.

aij the result of comparison between factor and factor , then we can construct pairwise comparison matrix .

Where aij=1/aji

Following this method, we can construct pairwise comparison matrix of criteria level and alternative level.

According to the meaning of quality of parts, we build pairwise comparison matrix of criteria level. The matrix is shown as follows:

Before we construct pairwise comparison matrix of alternative level, we first designate the level of the quality of each part and device, because the original content is described by words. The results are as follows:









According to this result, we construct pairwise comparison matrix of alternative level B1, B2, B3, B4, B5. The scale of these matrixes is too big, so we don’t display them here.

(3) Consistency Check

The definition of index of consistency check is CI=(λ-n)(n-1), λ is the eigenvalue of maximum of pairwise comparison matrix, n is the dimension of matrix.

The value of index of random consistency is as following:

We can obtain our desired RI from this table and calculate the ratio of consistency CR=CI/RI. If CR < 0.1, the degree of inconsistency is acceptable, otherwise, we need to reconstruct pairwise comparison matrix. After calculation, all pairwise matrixes accord to consistency checks.

(4) The Calculation of Parts’ Score

Supposing wA is the eigenvalue of maximum of pairwise comparison matrix A, w1 w2 w3 w4 and w5 is the eigenvalue of maximum of pairwise comparison matrix of B1, B2, B3, B4 and B5 respectively. Then the score of No.n part is:

The score represents the reliability of the part.

Then, according to the standard of “promoter-RBS-CDS-terminator”, we link all the parts together and construct a network, which is saved as adjacency matrix.

The weight of path is the reciprocal of the sum of two parts on the two sides of the path. As a result, when we search for the shortest path between the input and output, the path tends to go through parts with high scores, which indicates high reliability of circuit.

When users input the inducer and product of their desired circuit, we first determine the starting part and ending part according to users’ inputs, and then use Dijkstra algorithm to find out the shortest path between starting part and ending part. During the process of searching for shortest path, we make judgment at every step. The judgment is as follows: (1) the potential part must share at least one assembly standard with existing parts in the circuit, otherwise, the second shortest one is added to the circuit. (2) If there is more than one promoter in the circuit and the inducers and repressors of these promoters intersect, then we don’t choose this path, because the material that induces one promoter can repress the other, which can lower the rate of transcription.

Following the method above, we can calculate and display the circuit according to users’ input and show the score that indicates the reliability of the accomplished circuit. Users can also save the circuits in Nebula in case they want to check or change it later.




Future work

Now users can only share results as pictures, which is rather convenient for edit again. It will support other formats to share your design, for example, SBOL. What’s more, we will improve our interface to optimize the user experience. In version 2.0, all the parts in parts registry will be included in Nebula, which can greatly increase the convenience for users. Now we have realized the online version of Nebula with some basic functions, which is far from perfection. Much improvement will be made to perfect it in the near future and you can visit it on the page: http://cefg.cn/nebula/