9:20a – 7:00p, 11:00p-12:00a
Lab meeting
- final prep (9:20 – 10:00 A)
- presentation (10:10 – 12:30 P)
- lunch and journal club (Sang-Hee presents superesolution papers from last week’s issue of Science)
- includes Darzacq lab paper on polymerase
Meeting with XZ: notes
- Push ahead with cell culture
- put developmental applications on hold.
Project 2
- see post
Reviewing
- feedback for Sanchez/Golding
Scratch notes
Some of the data sets indicate whether the bursts measured are protein, mRNA or live (Fig 3), but at the moment the labeling is incomplete.
All 3 mammalian ones are just protein, (luciferase, and 2 GFP papers). Suter and Dar make use of the temporal information in infering bursts, I doesn’t appear to me that Skupsky uses live data despite the GFP based reporter.
I think Arjun’s 2006 paper provides a clearer estimate of bursting (for an entirely synthetic promoter, but Dar and Skupsky both use and HIV promoter so it’s not really an endogenous human gene either, even though it’s activity is also shaped by the chromatin context it lands in).
I think it’s a little confusing to compare protein ‘burst size’ and mRNA burst size side-by-side. I’m also more skeptical of the quantification of protein numbers and burst size. One can certainly measure variation in intensity. But without knowing absolute numbers you can’t say if it’s Poissonian noise or not (which means in fixed tissue one can’t easily quantify ‘burst size’. Dar et al measure live expression, so they can see the transcription bursts. (since it’s all the same reporter it is very unlikely the bursts are actually translation activity).
Skupsky 2010 looks at static protein levels. I don’t know what it means to compute a fano factor on arbitrary fluorescence units.
Suter et al also just use a luciferase reporter assay. All the plots labeled ‘number of mRNAs’, mRNA was ‘reconstructed from a mathematical model’. They never measure mRNA and there is no experiment in this paper that even attempts to validate the number they come up with for mRNA molecules based on their arbitrary fluorescence intensity values.
These papers can measure variation in burst size, but the data is of an entirely different nature (much less direct and in my opinion much less convincing) than other measurements Ido’s or Arjun’s.
A few of the protein papers monitor expression over time, so they can directly see bursts of different sizes. If one assumes the smallest observerable step is the protein created from a single mRNA, one can estimate burst size, though it would be better to show some data that this is true.
% 1 AU: 1 mRNA : 1 AU : 10 mRNA
[mean +/- std] 10 +/- sqrt(20) –> 100 +/- 10*sqrt(20)
Fano Factor: 20/10 = 2 | 2000/100 = 20
[mean +/- std] 40 +/- sqrt(60) –> 400 +/- 10*sqrt(60)
60/40 = 3/2 | 6000/400 = 60/4 = 15
(3/2) /2 = 3/4. (15/20) = 3/4