Sunday 03/02/14

Posted on March 2, 2014 by admin

10:00a – 11:40p

Today

  • working on journal club presentation (see googledocs)
  • working on team project: see post
Posted in Summaries | Tagged , | Comments Off on Sunday 03/02/14

Protected: team project update

Posted on March 2, 2014 by admin

This content is password protected. To view it please enter your password below:

Posted in Genomics | Tagged , | Comments Off on Protected: team project update

Thursday 02/27/14

Posted on February 27, 2014 by admin

10:00a – 11:15p

Meetings

  • (ski trip planning meeting) (10a-11a)
  • team project planning meeting (11a-12p)

Project 2 data analysis

  • locate new data files
  • run splitdax
  • run STORM analysis on data and beads
  • beads look low contrast but we’ll see.

Literature: selected for journal club

Notes on Lee … Church Science 2014: Highly Multiplexed Subcellular RNA sequencing in Situ

Concerning detection efficiency

  • intentionally exponentially deplete density using sequencing primers with random mismatches
  • on thresholds: use all signal (I suppose the brightest of the 4 channels is called the base?) relative rather than absolute thresholding? This could have been more explicitly explained than ‘putative sequences are determined for all pixels.’

Controls

  • induction of mCherry.
    • 7472 (98%) map to ‘+’ strand (as opposed to the minus strand? or as opposed to nothing? Given the method that every pixel is assigned a sequence, I suspect the claim of all reads mapped to on target can’t be true.)
    • shows nice induction from 0 to 15 transcripts per cell
    • For 853 genes from human primary fibroblasts with more than one observation (per cell?) compare to RNA-seq: Pearson’s correlation .57 fibroblasts to fibroblasts, .47 to lymphocyte seq, .23 to iPS cell seq. (excluding 1 outlier)

Experiment

  1. transcriptome in human primary fibroblasts
    • per-base error rate 0.64%, read 27 bases
    • ID 14,960 amplicons > 5 pixels -> 4,171 genes. (Do you get to choose which genes you look for? Does not seem so.
    • Fig S5 has ~250 amplicons implies of order 60 cells. 125-200 of these are rRNA. leaving 50-125 transcripts.
    • how stringent is the alignment ?!! Either you are missing most of the genes / sequences you could be detecting (by insisting on perfect alignment), or you are hiding an essential variable.
  2. trasncriptome primary fibroblasts after wounding. 171,730 reads, 6,880 genes.

Drawbacks

  • 81% of reads in EGF medium are rRNA

Goodbye Pubget

  • Pubget’s what’s hot in Science, top 10 papers this month: http://s2086772265.t.en25.com/e/es?s=2086772265&e=18778&elq=4ccdc583ab5d43e396c32c8bbb422845
  • includes a 1968 paper reporting the discovering of a restriction enzyme from E coli and a paper written in french on pallative care for the elderly.
  • I’m afraid pubget your filters don’t work for me.

STORM: imaging F04-647, F05-750

  • lower concentration (1uL primaries) seems a bit too low. Staining is clear but dots aren’t bright.
  • imaging buffer is 2uL fresh 1M MEA, 5 uL COT, 10 uL fresh GLox, 100 uL 50% glucose, 420 uL 200 mM Tris pH 8.0
  • we’ll see how things look in STORM
  • dye switching looks alright, density is pretty low (25,000 frames of switching instead of 50-80K).
Posted in Summaries | Comments Off on Thursday 02/27/14

Protected: project 2 planning meeting

Posted on February 26, 2014 by admin

This content is password protected. To view it please enter your password below:

Posted in Genomics | Comments Off on Protected: project 2 planning meeting

Wednesday 02/25/14

Posted on February 26, 2014 by admin

10:30a – 11:00p

Chromatin Project Analysis

Analysis of multi-color data

  • quick look at G1-G2 data from 12-30. Almost complete overlap of regions.

Current multi-color / adjacent region datasets:

  1. ‘Q:\2014-01-10_F11-7_F12-6’
  2. ‘Q:\2014-01-10_G01-6_G02-7’
  3. ‘Q:\2014-01-09_F05F06’
  4. ‘Q:\2013-12-30_G1G2’ – 750 survived remarkably well. Should check buffer on this
  5. ‘Q:\2013-12-28_F12F11’
  6. ‘Q:\2013-12-27_F12G01’
  7. ‘T:\2014-02-12_F05-6_F06-7’ – buffer died amazingly fast (in 1/2x BME?)
  8. ‘T:\2014-02-15_F06-6_F05-7’ – buffer died amazingly fast (in 1/2x BME?)

New staining

  • plate and fix cells
  • check cell density (just passaged yesterday, so this may not work).
  • density looks workable. Prepped for staining
  • new stains:
    • F06 A647 + F05 A750 (rpt, lower conc. hopefully better buffer)
    • F04 A647 + F05 A750
    • F03 A647 + F04 A750

code updates

  • added resolution option to STORMrender
  • substantial changes to ChromatinCropper in making it multicolor with new methods.
  • Ncolor still giving a little bit of trouble with colormaps when toggling active channels on and off.
    • when mlist is 1 dimension / 1 active channel the image is 2D and then Ncolor uses the colormap passed as the color range for the single color. So the 2nd active channel works fine but when only the first channel is active things screw up.
    • typically for multicolor images I’m passing Ncolor the hsv(2) or hsv(4) map. But for single color I use hot(256).
    • I’m sure I’ve made this somehow more complicated than it needs to be.

Meetings

  • project 2 planning (see notes)
Posted in Summaries | Tagged , | Comments Off on Wednesday 02/25/14

Tuesday 02/25/14

Posted on February 25, 2014 by admin

11:00a – 11:30p

Data compression (~11a-12p)

  • troubleshooting beaddax averaging
  • added some safety checks.
    • System now checks to see if detected files for deletion contain the downsampled ds tag. If they do they won’t be deleted.
    • System also checks if 561 target file to be deleted has a corresponding ds60 tagged downsampled file before deleting.
    • Added a try catch continue to bead-averaging. Rare movies give daxread errors for reasons I don’t yet understand. We just skip these and go on. The safety checks added above should preserve the original movie.
  • Code running smoothly, freeing up many TBs of data on ProBox.

Library 3 prep (~12p – 4p)

  • OligoArray troubleshooting
    • Want to kill BLAST jobs that hang, don’t kill java jobs, these will recover as you kill the BLAST.
    • Updated this in the OligoArray batch launcher function
    • This pipeline is getting smoother. Probably ready to start moving to its own functions.
  • Final checks for Library 3
    • anti-sense transcription checks out.
    • a few spot-checked probes align to correct targets with less than full 42 bp (e.g. 32). Seems to be a genome seq mismatch or something?
  • Ordering Library 3
    • Library design looks good, submit as Pending Order to get a PO. Then we send that to Marcelo along with the Fasta
    • converted space/tab separated primer data tables (Matlab, why is the default space-delimited instead of tab-delimited? Or LibreOffice at least thinks its just space delimited?)
    • fixed small bug in column headings
    • Submitted primer plates as IDT wishlist, Email to Alec, and entry on Pending Orders.

Deep Seq Error analysis (~4p – 6p)

  • using Jeff’s CIGAR parser
  • Overall looks pretty good
  • why do deletions increase steadily across the sequence body?
  • see post

Cell culture

  • passage cells (should restart from stock!)

Analysis of multi-color data (~6p – 11p)

  • high background 🙁
  • data-sets need organization
  • let’s tackle this more tomorrow
Posted in Summaries | Tagged , , , | Comments Off on Tuesday 02/25/14

Protected: Library Error Analysis by MiSeq

Posted on February 25, 2014 by admin

This content is password protected. To view it please enter your password below:

Posted in Genomics | Comments Off on Protected: Library Error Analysis by MiSeq

Monday 02/24/14

Posted on February 24, 2014 by admin

(sick, working from home)

Library3 prep continues

  • running primer finding
  • adding more filtering
    • do a direct comparison on final 5 basepairs,
    • reject all perfect matches, keep only unique ones.
  • reran a few times to debug
  • raising T min to 74C (76 gives too few primers)
    • this dramatically reduces the number of primers before BLAST.
    • BLAST runs way faster with thousands rather than 10s of thousands of primers
    • interestingly doesn’t dramatically reduce final primers out
  • Numbering is a bit annoying from mixed library runs
    • rerunning oligoArray overnight with new Lib3Table with all genes numbered

Ph Project

  • write to Ajaz about the FISH data
  • Ajaz also wants to know about the Psc data
Posted in Summaries | Tagged , | Comments Off on Monday 02/24/14

organizing research documentation

Posted on February 24, 2014 by admin

My Current Scheme:

Research

  • Papers
    • downloaded PDFs, a Mendeley ‘watched folder’
  • Paperwork
    • list of suppliers,
    • order forms for equipment,
    • reimbursement forms,
    • applications for grants / fellowships etc
  • Protocols
    • a github hosted repo of Markdown files
  • Presentations
    • conference talks / advisor chats / lab meeting talks
    • currently being moved to GoogleDocs Presentations
  • Projects
    • Docs
    • Results (stuff for figures)
    • Data (analyzed data files)
    • Code
      • Functions (project specific functions. e.g. parsing Kc data or polymer sim
      • Scripts (all playing around)
      • FigureCode (finalized script to produce particular figures on particular day)
  • Data (stored separately on drives)
    • folders organized by date:
    • YYYY-MM-DD_shortdescriptor

Problems

  • would like to have both private and public protocols
  • Scripts should be backed up github
    • (most are currently not on github)
  • Better function sharing
  • Make it more obvious where functions are
  • Repositories with published URLs
    • Image_Analysis (cited in Cell Reports 2013 and PNAS 2011 papers)
    • Shadow-Enhancers (cited in PNAS 2011 paper)

New Scheme

  • Protocols
    • Maybe these should be google-docs?
    • StackEdit.io can sync & modify the documents in google-docs
    • StackEdit.io can sync these documents to github for publication (the folder is just to create sub-folders in github).
    • see if we can link the google-docs for editing / private viewing of unpublished files
  • Code: New matlab-functions folder
    • contains lots of subfolders for very specific function tasks
    • BLAST
    • NextGenSeq
    • OligoArray
    • qPCR
Posted in Research Planning | Comments Off on organizing research documentation

dataset curation

Posted on February 23, 2014 by admin

Compressing Bead movies from the following data files

'M:\2013-11-01_F07\splitdax'
'M:\2013-11-02_D11\splitdax'
'M:\2013-10-29_E10\splitdax'
'M:\2013-10-28_E08\splitdax'
'M:\2013-10-27_E01\splitdax'
'K:\2013-10-26_D12\splitdax'
'K:\2013-10-25_D11\splitdax'
'K:\2013-10-11_F10\splitdax'
'K:\2013-10-10_F11\splitdax'
'J:\2013-10-06_no405Primary\splitdax'
'J:\2013-10-05_E06\splitdax'
'J:\2013-10-04_G05\splitdax'
'J:\2013-10-03_E02\splitdax'
'J:\2013-10-02_D09\splitdax'
'J:\2013-10-01_G10\splitdax'
'H:\2013-10-09_F01V2\splitdax'
'H:\2013-10-07_F09\splitdax'
'H:\2013-09-24_F6-405\splitdax'
'H:\2013-09-23_F6\splitdax'
'H:\2013-09-09_G4\splitdax'
'H:\2013-09-05_G6\splitdax'
'H:\2013-09-04_G3\splitdax'
'H:\2013-09-02_G2\splitdax'
'H:\2013-08-22_Ubx\splitdax'
'H:\2013-08-21_AbdA\splitdax'

All 561quad_movies downsampled to 1 Hz, introduce gain of 4 to keep some dynamic range of data, original 60 Hz movie deleted. (Gain of 10 would saturate some frames of some movies).

Need STORM split-dax

  • J:\2014-01-08_BXC

Need 2 color STORM split-dax

  • J:\2014-01-09_PhM (647 and 750)

Need daoSTORM fitting

  • H:\2014-01-29_F03F04 (647)
  • J:\2014-01-09_PhM (647 and 750)

Has issues

  • H:\2014-01-12_G04 (mislabeled daxfiles?) check this!
Posted in Chromatin | Tagged | Comments Off on dataset curation