9:45a – 11:30p
cell staining
- heating buffer.
- wrong buffer, no formamide in 60C hot wash.
- Did I use the formamide buffer yesterday? I recall heating a 50mL tube and I don’t think I have 2x SSCT in a 50mL tube… maybe this was the origin of the loss of staining…?
- hot wash, 20 min, 60C, 2x SSCT
- check staining: background looks high, dots kinda weak (much fainter than BX-C for sure), but very good switching. Probably a background/contrast problem. Could use less probe.
- STORM imaging looks pretty good.
- 405 probably still too concentrated (try diluting another 1:10 before use, compare to a no-405 control).
- background pretty high, could probably reduce probe concentration substantially.
BXC FISH probes next steps
- Add feducials. Try both cross-linkable 415 beads and non-cross-linkable 514 beads?
- score coverslip — see if we can find the same cells again. Try a hatch pattern with the diamond etcher to start
Review
- editing intro.
- working on abstract.
- rewrote abstract.
- editing discussion.
- got copyright permission to reuse Raj image of worm embryos from 2010 Nature paper.
Ph project
- data analysis
new approach:
- bin localizations
- toss isolated bins with small number localizations
- cluster remaining localizations
- clusters over maxSingle threshold are clusters of localizations, others are considered dispersed/non-clustered localizations.
- look at all Ph-flag localizations. Pair with nearest Pc localization cluster within some min centroid-centroid distance (e.g. 150 nm. Should probably due cluster edge or cluster overlap or some other metric).
- determine fraction of those Pc partners that are in Pc clusters (relative to all paired Pc localizations).
- ‘+’ indicate isolated localizations, ‘.’ indicate centroids of clusters.
Preliminary results on cell scale
fracNotClust =
0.0331
0.1541
WT: fraction Pc in clusters near Ph localizations = 0.7442
fracNotClust =
0.0808
0.1265
WT: fraction Pc in clusters near Ph localizations = 0.5723
* note that no substantial difference in overall fraction clustered, (goes either way), but more consistent difference in fraction of Pc clustered when using Ph-flag as filter.
STORM
- O/N STORM on scored section of cells
- hopefully we can find this score again.
- collecting 512×512 movies with quadview, tracking beads in one pannel and cells in other. (should definitely write a 256×512 version), current approach is very wasteful of data-space!