||The model is parameterized using mRNA expression observed by Weiner et al for M. pneumoniae at 37°C [PUB_0569]. Additionally the knowledge base includes Weiner et al's cold shock (32°C) and heat shock (43°C) gene expression measurements. Guell et al have performed 62 additional gene expression measurements of M. pneumoniae across a variety of experimental conditions [PUB_0418].
Since Weiner et al did not provide a complete table of expression means, expression data was inferred from Figure 4: an expression map of M. pneumoniae where a color scale represented normalized expression. The color scale ranged from -3 to 3 with 10 intervals. The color of each gene was recorded for cells grown at 37°C. These colors were then compared to the numerical data provided for a subset of genes in Tables 1 and 2. From this, a Matlab script was written to find the optimal numerical bounds of the color scale intervals such that a maximum number of provided numerical data fits the bounds. Constraints were that the intervals had to lie between -3.3 and 3.3 and that the intervals had to be of equal size. The optimized intervals were as follows:
-2.945 to -2.3421 (lowest expression)
-2.342 to -1.7391
-1.739 to -1.1361
-1.136 to -0.5331
-0.533 to 0.0699
0.070 to 0.6729
0.673 to 1.2759
1.276 to 1.8789
1.879 to 2.4819
2.482 to 3.0849 (Highest expression)
Then for each gene for which numerical data was not provided, the average value of the bounds of its color scale interval was assigned. Their data was normalized by subtracting the mean background from the raw data and taking the log of the result. From this, the average membrane signal was subtracted and the result was divided by the standard error. Since the mean background, average signal, and standard error were unknown to us, to re-derive expression levels we simply calculated 2^(color scale value).
This gave us relative expression values for most of the M. pneumoniae genes, which we assigned to their corresponding M. genitalium genes from the M. pneumoniae-M. genitalium mapping.
The total mRNA “expression” (in arbitrary units) was defined as the sum of all genes’ mRNA microarray data defined above as 2^(color scale value). We made the assumption that the ratio of total mRNA “expression” to the number of total mRNA molecules would equal the ratio of total tRNA “expression” to the number of total tRNA molecules. This gave us the total tRNA expression. To get the expression for each individual tRNA, we took the fraction of total residue dry weight per tRNA of this value.
rRNA expression was determined similarly to tRNA expression. As described above we know the number of molecules of each type of rRNA in the cell. Given this, and the assumption that the ratio of total mRNA “expression” to the number of total mRNA molecules would equal the ratio of a given type of rRNA “expression” to the number of that type of rRNA molecules, we obtained estimates for the “expression” of each type of rRNA molecule.
Güell M, van Noort V, Yus E, Chen WH, Leigh-Bell J, Michalodimitrakis K, Yamada T, Arumugam M, Doerks T, Kühner S, Rode M, Suyama M, Schmidt S, Gavin AC, Bork P, Serrano L. Transcriptome complexity in a genome-reduced bacterium. Science 326, 1268-71 (2009). WholeCell: PUB_0418, PubMed: 19965477
Weiner J 3rd, Zimmerman CU, Göhlmann HW, Herrmann R. Transcription profiles of the bacterium Mycoplasma pneumoniae grown at different temperatures. Nucleic Acids Res 31, 6306-20 (2003). WholeCell: PUB_0569, PubMed: 14576319