Team:Cambridge-JIC/Informatics

From 2014.igem.org

(Difference between revisions)

Revision as of 17:14, 20 August 2014

Informatics

To introduce a novel chassis and make full advantage of its properties, we want to computationally analyse its genome in order to:

find out the most efficient codon usage in order to optimise our registry parts and facilitate all future synthetic biology work on Marchantia;
identify and characterise promoters, in particular looking for strong, inducible, tissue-specific or early development stage promoters. These are valuable tools to further manipulation

Codon usage optimisation

Codon Optimisation

Our start point was the Marchantia genome and the mRNA transcriptome predicted with the Geneious software (http://www.geneious.com/). The data was given to us from Jim's lab by Bernardo. 99 000 ORFs were predicted, which seems too large to be realistic. Half of these were only 100 aa long. We set the threshold for candidate genes amongst these at 300 aa, obtaining the expected normal distribution of lengths.

Using the BLAST software (http://blast.ncbi.nlm.nih.gov/Blast.cgi), we compared the proteins coded by these candidate genes with the proteins present in Arabidopsis, given on Araport (www.araport.org). Some of the sequences showed incomplete matches, indicated that our predicted ORFs should be regarded with some vigilance. The longest sequence showed a 40% match, a small number as expected.

The relevant DNA sequences from the candidate mRNAs using the BLAST output. Then the different codons were counted in these genes, in order to obtain a codon table for Marchantia. While the table is not strikingly similar to Arabidopsis', we can note a similarity in the slight preference for C over other bases at the end of codons and that for G-p-C sites.

A .xls version of this table with relevant filters can be found here.

Promoter identification

To identify potential promoter regions in Marchantia's genome, we looked for gene sequences from other species, mainly Arabidopsis thaliana and Physcomitrella patens. We were particularly interested in genes that were:

Nitrate inducible
Sulphate inducible
Phosphate inducible
Circadian rythm driven
Metabolism related
Development related
Light responsive
Development
Standard templates for proteins
Standard templates for genes

Reading recent research papers in search for these genes and looking for their sequences on GenBank and ThaleMine, we constructed a list of protein sequences to compare to the Marchantia predicted scaffolds. (For genes where only the nucleotide sequence was available, instead of the protein sequence, a C++ code was written to perform the translation). We ran a tblastn search in BLAST for the best matches between our candidate proteins and the Marchantia scaffolds. Then, for the best matches (~60% and above, with some judging by eye), we annotated the 2kb upstream of the start codon as a potential promoter region.

We identified 30 candidate promoters this way, that we are planning to screen by inserting in a construct driving the yellow fluorescent protein Venus. For each promoter, we will make a construct with and one without amplification by GAL4 and GAL4 UAS, to evaluate the promoter strength and get around any leakages due to the use of GAL4.

@@ Line 23: / Line 23: @@
 <h3> Codon usage optimisation </h3>
+<a id="https://2014.igem.org/Team:Cambridge-JIC/Informatics#Codon-Optimisation">Codon Optimisation</a>
 <p>
 Our start point was the Marchantia genome and the mRNA transcriptome predicted with the Geneious software (<a href = "http://www.geneious.com/">http://www.geneious.com/</a>). The data was given to us from Jim's lab by Bernardo. 99 000 ORFs were predicted, which seems too large to be realistic. Half of these were only 100 aa long. We set the threshold for candidate genes amongst these at 300 aa, obtaining the expected normal distribution of lengths.