Team:DTU-Denmark/Achievements/Interlab study

Construct strains

We created multiple strains expressing GFP from different promoters from the Andersen Library. 12 out of the 15 promoters we intended to use were successfully transformed into DH5α These 12 constructed strains were applied in the further analysis of the relatively promoter strength. The promoters J23111, J23117 J23109 was not successfully cloned into E. coli. We did not detect any fluorescence signal from these strains and our sequencing results confirmed that they were not correctly constructed. These 3 strains were excluded from the dataset. To measure the background signal the E0240 was used with GFP without any promoter. The I20260 BioBrick was also applied. I20260 also contain the J23101 promoter but present at another backbone than the remaining promoters.

Fluorescence measurement on cultures

Fluorescence was measured on the O/N cultures with our constructed strains in the BioLector. Biomass and fluorescence were measured during growth. Fluorescence signal through the growth was normalized by dividing by OD600. The average of the fluorescence is illustrated in the bar chard below. Together with the expected values relative to the J23100 promoter. Orange is the expected values and gray indicates our measured fluorescence signal. THe values indicated in the bar chart are mean values of all the measurements taken during a 3-hour period, with 4.88 min intervals. For more details on the measured values and the complete dataset look at the interlab form.

The data was analyzed by fitting a statistical mixed model to the data, using the lmerTest package in R. The different promoters were modeled as a fixed effect, and replicates were modeled as a random effect. The estimated means for each promoter and the standard error of the mean can be seen in the chart below.

In two of the triplicate sets, one outlier was excluded: For the promoter J23100, one replicate did not grow very much. For promoter J23102 one replicate had negligible fluorescence.
The background-subtracted values were calculated by subtracting the mean of E0240 from the other means. The replicate standard deviation was estimated to be 0.014, meaning that the effect of using multiple replicates contributes a random effect distributed with this standard deviation. The residual standard deviation was estimated to be 0.040, meaning that the random effect of taking multiple measurements on the same sample has this standard deviation.
From the data we conclude that many off the evaluated promoter activities are to some extend in agreement with the expected activity. However it seems that the weak promoters are rather difficult to measure in the BioLector.

This interlab study is a contribution to the bigger research collaboration. It will be interesting to see contributions from the other labs and whether these are in agreement with our obtained data.