lrgr / explosig Goto Github PK

Buses are nicer than watchers ($emit, $on)
see https://alligator.io/vuejs/global-event-bus/

https://www.nature.com/articles/nature12477/figures/1
mutations per megabase
deconstruct by mutation category or mutation base substitution type

prevent dispatching,etc when using local files

Rearrangements are hypothesized to associate with kataegis. It would be interesting to add rearrangement "breakpoints" to the rainfall plot.

To get the breakpoints, you would structural_somatic_mutation files, e.g. for BRCA-EU at this link:
https://dcc.icgc.org/releases/release_26/Projects/BRCA-EU

No requests should specify chromosomes, just fetch data for all, then show/hide on front end

will require adding svg xmlns attributes everywhere
and figuring out weird axis tick text positioning

Same process as minimum bar width for exposures plot

Allow user to drag along a bar above/below genome-wide plots to show a different chromosome
Possibly just never limit to single chromosome, always load data for all (kind of already done for kataegis plot)
If so, create plots for all chromosomes, with only a single one visible initially (think of as window)

Plots that use aggregated data

• manhattan plot(s)
• avg exposure levels?
• total exposure levels?
• clinical data stats?
• selected signatures

• 560 BRCA from ICGC
• Kasar, et al. (Nature Communications, 2015)
• The two non-TCGA datasets from Kim, et al. (Nature Genetics, 2016)

non-COSMIC signatures:

• Nik-Zainal, et al. (Mutagensis, 2015): https://doi.org/10.1093/mutage/gev073
• Kasar, et al. (Nature Communications, 2015): http://rdcu.be/IRAv
• Kim, et al. (Nature Genetics, 2016): http://rdcu.be/IRBd
• Chang, et al. (Nature Communications, 2017): http://rdcu.be/IRA5
• Supek & Lehner (Cell, 2017): http://bit.ly/2IhYeFI
• Hanninen, et al. (PLoS Genetics, 2018): http://bit.ly/2Inz8Wg
• Alexandrov, et al. (bioRxiv preprint 2018): https://www.biorxiv.org/content/early/2018/05/15/322859.full.pdf+html

Give each signature a name, a source - publication, and an index
Sort by publication, then index

http://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/ possibly use refFlat.txt.gz file

exons, transcripts, promoter regions?

possibly see http://epiviz.cbcb.umd.edu/4/

Choose a different set of signatures, plot differences in exposures

Not separate for each plot

Look into using VAF to visualize clonality
Will probably need to retain duplicate donor reads/samples to do so
Think about this more
https://www.biorxiv.org/content/early/2018/02/05/260471

• rainfall plot context colors
• exposures plot signature colors
• cancer type/project colors
• exposure plot clinical data boxes

Local file reading works in Firefox but not chrome

https://github.com/Nanonid/rison

https://bl.ocks.org/mbostock/5872848
https://github.com/d3/d3-dispatch

See how MAGI implements this http://magi.brown.edu/

limit redundant code

• exposures plot

Relates to #34

not urgent, but would be nice to work in at least:

• firefox
• chrome
• safari (including mobile)

Upon addition of new plot, show small animation to indicate where dragging functionality is enabled

see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5477513/ supplement 1 pg 4
number of mutations per sample, by signature exposure, comparing smokers/non-smokers and <60/>=60 yr olds

https://github.com/d3/d3-drag

• tobacco
• tobacco intensity
• alcohol
• sex
• diagnosis_age
• diagnosis_icd10
• vital_status
• prior_malignancy
• num_relatives_with_cancer_history
• first_therapy_type
• first_therapy_reponse
• second_therapy_type
• second_therapy_reponse

save whether single donor or all donor mode

very minor