Saturday 12/28/13

9:30a – 1:45a

Library 3 Design / probe building

oligoArrayCommand was getting rewritten in the loop, need a temp parameter to preserve the genename flag to be used for substitution
consequently probe building on both Cajal and Monet failed.
Cajal returns all the same missed BLAST hits.
- I wonder if this is do to flipping genes to sense sequence.
- many probes hit nothing, some probes hit the correct chromosome, some probes hit the correct and incorrect chromosome, some hit only the incorrect
- this makes me wonder if oligoArray is too inflexible to BLAST complementary sequences. I suppose I could write a short direct test of this.
fixed temp parameter error, rerunning on Monet and Cajal (with respective Dmel_Genome (single fasta) and Dm_Genome – (fasta by chromosome)
when running on a single machine with multiple processors it is just as fast to run more parallel processes (especially for large regions that are split into many sub-regions. If the average gene size is 1kb this method is less useful. But when its 100s of kb this is probably simpler.
When blastall hangs the whole process freezes. Finding and killing the frozen blastall resumes analysis. (If no blastall it might be the conhost.exe that freezes). Not sure yet what this does to the corresponding oligo-file.

Confirmed: Oligo-Array is strand specific.
I suppose we can flip the genome and make a bigger fasta file that has each chromosome and its reverse complement.
caught another error — we really do need to seqrcomplement not just seqcomplement to get the coding sequence on the minus strand. Also note we will need to seqrcomplement again in building the probe library, since we take the sense strand and we want the probes to be the antisense so they can hybridize to the RNA.

testing higher temperature cutoffs on Cajal. Especially for short sequences more probes would be nice. I think this is actually why the piecewise assembly via nicked translation of short regions works better. The piecewise assembly part helps balance relative probe concentrations, and the very short overall segments of long probe lengths (often in the 100 – 300 bp range I believe, studed with dyes) ensure substantial signal is localized to the target.
Array sensitivity really does come from a single sequence (with multiple copies of it per well/pixel). But in our method single probes are pretty meaningless, and we can be much more robust to differences.

Updated overlay management with new button structure
Added chromatic warp options to overlays (finally).
started working on Tutorial for STORMrender
fixed sliders.
damn, didn’t save all files before switching branch, windows doing crazy denied permission bs.

F12-647 pretty weak on F12-647+F11-750 double stain.
F11-750 stains look excellent!
concern: some cells have 6 or 7 copies of this locus. (I guess the cell is tetraploid and if the sister chromatids in G2 come apart we expect 8).
750 switches extremely well (really need to back down the 405)
647 bold and strong. See how the buffer run-down goes
The newly prepared, newly adhered beads on this particular sample are not bleaching nearly as badly as before.
COT (I think that’s the weak point at least), substantially run down by 620.

Glass PDMS seal not so good for pulling flow through. System likes to pull in airbubbles through the glass PDMS interphase to relieve the pressure.

13 samples
14*1.7 ~ 24 uL of each rNTP + 26 uL T7 buffer + 28 uL T7-polymerase + 14 uL RNaseOUT = mastermix.
11 uL of this to each (9 uL) sample of template DNA.

stains look horrible. Cy7 nicely outlines the nucleus.
Dramatic cell-cell variation in intensity of native Ph as well as Ph-flag. Not particularly correlated.
Try new permiablization procedure?
Try new antibodies?
stain controls for transfection efficiency?