Alistair's Lab Notebook

9:45 am

Chromatin Paper

• writing first draft of letter to editor
• working on probe table
• still need to make primer table.

MERFISH

• working on revised analysis pipeline

Personal stuff

• looking at apartments

11:00 am – 1:00 am

Chromatin manuscript

• computed p-values for overlap figures. Added values to fig caption
  • maybe should go into main text when I get it back from Jeff
• p-values for overlap fraction for reference
  • Repressed-Active interact less than Silent-Active, p < 1e-14.
  • repressed intradomain interact more than active, p < 1e-26.
  • repressed intradomain interactions more than silent, p < 1e-22.
  • silent intradomain interact more than active, p < 1e-2.
  • silent active interact less than silent-silent, p < 1e-10.
• p-values for entanglement score for reference
  • Repressed-Active interact less than Silent-Active, p < 1e-4 (~1e-5)
  • repressed intradomain interact more than active, p < 1e-12.
  • repressed intradomain interactions more than silent, p < 1e-9.
  • silent intradomain interact more than active, p = .1 (Not signficant)
  • silent active interact less than silent-silent, p < 1e-4
• updated linker colors in Extended Data Fig 6
• computing p-values for for extended data fig 8

9:30 am – 11:55 pm

New Data analysis tools

• generalized Hi-C plotter with Bogdan
• now does triangles too
• uploaded new HiC data. Looks much smoother, replicates with different enzymes look very similar.

Chromatin paper

• working on manuscript revisions
• discussed plans and progress with LM and team (6pm-8:30pm)

To do still

• comment on cell cycle and pairing in intro section (may need additional supplemental discussion)
• add p-values to all comparisons in Fig 2, 3, and Extended Data Fig 8 (and 6 and 7), with N cells per comparison.
  • XZ recommends combining samples of same type for the fig 2/3/6/7 comparisons.
• check entanglement scores (rerun?)
• fix connectors
• combined fasta file

9:00 am – 8:30 pm

morning 9 – 10:45am

• chat with HC
• email Geoff, congrats and follow up
• email DB about potential collaboration
• create 1 page pdf of recent publication info for K99 update

Chromatin paper methods revision

• working on description of domain selection, 10:45 am – ?

9:00 am – 10:15 pm

Geoff’s Defense (9:30 am – 11:00 am)

• see post
• short follow-up discussion with BB and GN

To Do

• Travel reimbursements
• write to XZ about LM travel schedule (done)

Chromatin Paper

• Built new composite data structure (allData.matb) to avoid mixed data sets and crazy indexing.
  • some manual field alignment: print elements to Excel, ID missing fields, annotate back in by hand.
  • built new domainData2 structure which has the library number and positions (just positions was getting very confusing)
• Completed assembly of Supplemental Figure on Intact Domain Histograms

Extended data Figure revisions (3pm – 5pm)

• renumbering extended data figures in Main text, figures, and figure captions.
• Writing new extended data figure captions.

New Extended Data Figure 8:

• wrote generic bootstrapping function to compute confidence intervals.
• set confidence intervals to 78% for errorbars, So probability two data points with confidence intervals that just overlap are not different ~.22^2 = 0.05.
• alternatively could do the traditional SEM and the error bars would be smaller still.

Still to do

• Write cover letter
• Write description of domain selection
• color connecting bars in Fig E6
• print multiple versions of green/magenta/white contrast for Fig 2 and 3
• combine probe fasta tables and update probe names with new conventions

MERFISH-lunch

• Jeff will send script on new probe design for human genome
• Try building a matlab class to handle library design?

Chromatin Paper

• updated PlotGenes to fix issues with arrowheads. Much nicer now.
  • applied in Extended Data Fig 3. Others still need updating.
• Working on Extended Data Fig 6: illustrating differences between active domains of similar size.

Paper To Do list

Text

• For the abstract and mostly intro: don’t over-emphasize epigenetic structure. We study structure at the relevant length scales for gene regulation / genome functions. Epingenetic domains are one of the indicators that this length scale is important, and it turns out a strong predictor of structure.
• With respect to Blue internal scaling, can’t say most models — invites question of other models. Move to discussion of models, not in data. Discuss the equilibrium globule model. Say no model fits all the data for Blue.
• Need to discuss off-trend points in model section
• be sure word “loop” does not appear in text
• need to write Cover Letter.

Main Figures

• add scalebars to Fig 1a
• need multiple examples of different contrast /saturation options for Fig 2a and b
  • contrast / saturation for Repressed domains is a bit strong.
  • images are a little blury — plot with different Gaussian widths

Extended Data Figure

• New Fig
  • histograms of volume and radius of gyration distributions for all domains.
  • similar histograms for FG model data?
• ED2 ‘internal scaling of repressed regions’
  • reorder panels. BX-C vol, ANTC chip, ANTC rg, ANTC vol
  • All int data plots: solid lines for int-data, dashed for old data
• Tracing fig
  • no backbone
  • 15 spots
  • green-green junctions linked by green lines. magenta by magenta lines. Black lines between interjunctions
  • also in data follow junction code.
• New example stats
  • select regions similar length
  • show existing stats
  • also show cell-cell variation

9:00 am – 9:50 pm

Chromatin Manuscript

• revising figures and color balance
• helping XZ images and slides

Other stuff

• discussed projects with BB
• discussed L12 progress with JM and HC
• helping GN troubleshoot dax readers