- AP patterning (13)
- Blog (1)
- Chromatin (85)
- Conference Notes (48)
- Fly Work (54)
- General STORM (25)
- Genomics (130)
- Journal Club (19)
- Lab Meeting (51)
- Microscopy (78)
- Notes (1)
- probe and plasmid building (58)
- Project Meeting (3)
- Protocols (13)
- Research Planning (70)
- Seminars (17)
- Shadow Enhancers (59)
- snail patterning (40)
- Software Development (4)
- Summaries (1,222)
- Teaching (9)
- Transcription Modeling (40)
- Uncategorized (10)
- Web development (19)
Tagsanalysis cell culture cell labeling chromatin cloning coding communication confocal data analysis embryo collection embryo labeling figures fly work genomics hb image analysis image processing images in situs Library2 literature making antibodies matlab-storm meetings modeling MP12 mRNA counting Ph planning presentation probe making project 2 project2 result results sectioning section staining shadow enhancers sna snail staining STORM STORM analysis troubleshooting writing
Warning: Invalid argument supplied for foreach() in /home/content/06/7103106/html/wordpress/wp-content/plugins/latest-github-commits/latest-github-commits.php on line 113
9:30 am – 10:15 pm
- reply to Ajaz and Nicole about modeling work
- send back comments on plan to add model figure
- should try to compute fraction of localizations/molecules in clusters over 30 nm as a function of total Ph concentration for all types. Try this later…
Oligo Secondaries project with Wu lab
- final comments for last round review / response to reviewers (hopefully) sent to BB
Chromatin scaling manuscript
- working on introduction
Discussions on HiC
- Questions for Max and Geoff:
- evidence of domain switching (looking at the large scale, off diagonal interaction preferences, change a domain from repressed to active (in a different cell line or by stimulation) and show that it switches from compartment B to compartment A).
- supposedly this is clear from Dixon 2012 and Nora 2012 data
- 3C / 4C / 5C hard to have a simple and accurate normalization — not clear what’s visibility vs. contacts
- linker domains hard to formalize
- domains follow epigenetic marks but not dependent on them
- CTCF / cohesin not yet shown to have a role in metazoan genomes but maybe that’s an assay problem: sufficient time for domains to re-equilibrate after knockdown? sufficient knockdown?
- supposedly removing a CTCF (or cohesin) loader is sufficient to cause more substantial changes?
improving 3D data
- looking for better correction factor (see RecomputeChromatinVolumes_150330).
- simple attempts at linear rescaling based on radius of gyration, computed volume, rg ratio, don’t suggest that such an approach is promising.
- not sure the z-data is quantitatively accurate enough for these measurements.
- blue data pretty reliably still for large regions in all dimensions, but for large black and large yellow there is a systematic depletion of the data at the z-extremes due to detection efficiency issues, which causes these trends to curve down. Images look okay but quantification not as good.
9:00 am – ?
Getting Drosophila expression data
- Failed to find assembled FPKM data
- downloading 0-4hr data sets from modENCODE
- What’s with the big time window. There must be better these?
- need a GTF file for cufflinks to parse the SAM data
- running cufflinks: errors
- need to convert SAM to BAM
- downloaded GTF from UCSC refFlat.txt – this gives errors
- downloaded R5 data from ENSEMBL (release 5 is getting hard to find now everything defaults to release 6) ftp://ftp.ensembl.org/pub/release-77/gtf/drosophila_melanogaster
- RC Odyssey is very slow today. Was giving lots of acct errors too: server not responding. And my jobs stay in Pending purgatory for an hour. Better than all day I suppose.
- finished design of probes, check probe length
- still have hundreds with good length
- computed sum FPKM from slicing data (this data is truly blastoderm — Hox genes and Segment polarity genes inv / en are cold off. Would be lovely to have some later embryos. Shame the modENCODE data is so broad. How about some single embryos at close timepoints?
- downloaded blastoderm single embryo data from Susan Lott’s paper
Ordering control probes
- submitted orders to Alec for BioSearch / Stellaris probes for Kr and tll as positive controls.
9:00 am – 6:00 pm
- discussion with patent lawyer
- discussion with MERFISH team
- User Profile Disks might be problem with user session management: http://www.virtualizationadmin.com/articles-tutorials/vdi-articles/general/working-with-user-profile-disks-on-session-based-desktop-deployments.html
- disabled User Profile Disks (these were not enabled on CAJAL but were enabled on MORGAN when MORGAN started deleting user account data).
- re-installing local code database on my MORGAN account
- running OligoArray with 20mers 66C – 76C for all IMR90 transcriptome (37,442 genes)
- launches about as fast as when running on Cajal, but uses no CPU (maybe saving over the network uses more CPU?)
Tissue section prep
- Ultratome booked today and tomorrow. I booked Friday after lab meeting
- working on slides for journal-club presentation on past 2 years of Hi-C.
9:00 am – 6:00 pm
- Gomez-Diaz and Corces 3D genome architecture review
Drosophila Patterning screen
- started new Projects folder TF interactions
- Downloaded list of 1000 annotated sequence-specific TFs (GO annotation) from FlyBase. This list has substantial number of non-melanogaster genes to be removed.
- Extract gene IDs, Extract sequences
- Flybase exon download extracts exons of different lengths but mostly overlapping sequence as different exons.
- use transcripts instead, just select longest transcript
- order tiling probes from BioSearch (30 probes) against two select genes
Potential cover art ideas
- move embryos into 30% sucrose for cutting
9:00 am – 7:30 pm
Finalizing draft of chromatin manuscript
- revised gene expression panels in figure 1
- added additional references to text
- wrote figure captions
Additional references for chromatin literature
- chromatin organization and function (spatial positioning effects) 2011 http://www.sciencedirect.com/science/article/pii/S0955067411000196
- chromatin organization and function (editorial, histone code and 3D genome organization) 2008 http://ac.els-cdn.com/S0027510708002522/1-s2.0-S0027510708002522-main.pdf?_tid=abb4d7a8-d194-11e4-906b-00000aacb361&acdnat=1427139899_12a077eb9fafec4dbe28b479c17def4b
- Insulators review (2003): http://www.sciencedirect.com/science/article/pii/S0955067403000395#
- Condensins and chromatin organization: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3927828/pdf/nihms528141.pdf
- Nucleus (editorial intro) http://www.sciencedirect.com/science/article/pii/S0955067414000581#. Discusses 3D genome organization hypothesis
Other relevant chromatin literature
- continue reading up recent Hi-C publications:
- van Bortle 2014 Genome Biology on Insulator function
- Seitan 2013 on Cohesin
- Schoenfelder 2015 on Promoter-enhancer interactions by promoter capture Hi-C
10:00 am – 12:45 am
FindDataFiles to return a list of all the folders on data drives. These can be passed to StringFind to find
- finalizing layout
- added gene expression bar-graphs (doesn’t look that good)
- added quantifications
- explored alternative panel layout
- modified internal scaling to move legends outside. Still need to update printing of these.
- wrote new code to process new domain data.
- manually rescreened all data. saved as new interactData.matb in the Index folder in Chromatin\Data
- minor screwup saving frac_irINred and frac_redINir into interactData structure
- not a real problem as these can be recomputed from the raw data also saved in dist_ir2red and v.v. See Fig 4 assemble script.
- started new version of the figure script. cleaned up the code a little bit.
- fixed schematic images as discussed.
- very short black inside a very large polymer don’t coil back on self enough. Need larger black. Trying this now.
- Figures probably in good enough shape to return to text…
9:00 am – 11:30 pm
- re-analyzed C02 data (originally taken on STORM1, not the best data, background separation a bit difficult).
- looking an Kc CNV
- this genome is a bit of a mess (not sure how much noise is in this data either), but spot checking regions none seem especially or broadly out of average. Bxd is apparently over-replicated.
- tried to compare to clone8, but the modEncode link is broken
File names explain metrics (click to see). See also y-axis label
- CoV of pdist matrix
- this looks promising:
9:15 am – 7:00 pm
- analyzed up through dot 2 (29) of new D08 data (through dot 64). Looks great.
- analyzing fractal dimension parameters — computing for all composite Flists (this is very slow so hopefully I’m computing it right!)
- laid out fig4