First great battle in the war on noise

Dear readers, I hang my head in shame for a glaring omission. At the end of 2015 Nature Methods published: Quantitative characterization of genetic parts and circuits for plant synthetic biology. This work was a collaboration between June Medford’s plant synbio lab and Ashok Prasad’s Engineering lab, both at Colorado state.

The lofty aim of that study was to characterise 128 different combinations of synthetic repressor TFs (Gal4 or LexA DNA binding domain + one of four plant repressor domains) and repressible promoters (diff. consitutive promoters with cis-elements for the repressors in different positions). However, a quick reading of the paper is sufficient to see that the real focus of the study turns out to be not the activity of the repressor-promoter pairs, but the issue of biological noise.

Most of the work was carried out in Arabidopsis leaf protoplasts and they found that noise arising from the luciferase measurement procedure and different protoplast batches drowned out the effect of the different constructs. Whilst they came up with a correction method for the luciferase readings (I wont even pretend to understand the details but they are in abundance in the supplement for those interested), this was not so easy for the batch effect. The first step was to identify the problem as arising from a batch effect, as opposed to noise coming from individual technical replicates (that effect was there but less important). Then they had to work out how to calculate a batch factor and how to apply this factor to their luminescence reads data to give the best normalization.

So far so good. However, after going through all this they compared expression levels from protoplasts transiently transfected to those from stable transformants containing the same reporter+repressor pairs. This is found in figure 6, which I include here in the hopes that Nature don’t sue me:


The authors were at pains to point out that “For all three parameters, the error bars of the estimates typically overlapped, and for these estimates the differences were found not to be significant (P = 0.1).”. But the trends between the three different promoter-repressor pairs don’t even match between the transient and stable samples. This reminded me of a figure from a study out of John Dueber’s lab characterizing a toolkit of gene expression parts in yeast:


The take home from Lee et al was clear – if you want reliable data use genomic integration. Sadly I’m not even sure if something so clear can be said for the plant study. As all plant molecular biologists know, the expression of transgenes derived from Agro transformation can differ wildly, presumably due to the effect of genomic integration site.

Whilst I can’t see that Schaumberg et al., provides any clear take homes on synthetic promoter design in planta, it does provide a rallying call for better tools for the plant synbio community. What we need are easier tools to create stable transgenics with insertions at defined AND CHARACTERIZED genomic locations. Considering the rise and rise of gene targeting technologies this seems eminently feasible and I wouldn’t be too surprised if pretty soon genomic integration at defined loci is the norm, as in yeast synbio. The virus-T-DNA transformation method from the Voytas lab is certainly a step in this direction and I will keep you posted on further developments. Obviously things slip through my net from time to time – so if you see something that belongs on this blog PLEASE EMAIL ME –






Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s