Home

Last updated: 2018-08-22

workflowr checks: (Click a bullet for more information)

✔ R Markdown file: up-to-date

Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.
✔ Repository version: ac09d4b
Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility. The version displayed above was the version of the Git repository at the time these results were generated.

Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:
```
Ignored files:
    Ignored:    .DS_Store
    Ignored:    .Rhistory
    Ignored:    .Rproj.user/
    Ignored:    docs/.DS_Store
    Ignored:    docs/figure/.DS_Store

Untracked files:
    Untracked:  data/greedy19.rds
```
Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.

Expand here to see past versions:

File	Version	Author	Date	Message
Rmd	ac09d4b	Jason Willwerscheid	2018-08-22	wflow_publish(“analysis/index.Rmd”)
html	b0fbbe0	Jason Willwerscheid	2018-08-22	Build site.
Rmd	4fe662c	Jason Willwerscheid	2018-08-22	wflow_publish(“analysis/index.Rmd”)
html	02be9e9	Jason Willwerscheid	2018-08-17	Build site.
Rmd	6e0036f	Jason Willwerscheid	2018-08-17	wflow_publish(“analysis/index.Rmd”)
html	a6d55b6	Jason Willwerscheid	2018-08-16	Build site.
Rmd	c972927	Jason Willwerscheid	2018-08-16	wflow_publish(c(“analysis/index.Rmd”, “analysis/parallel2.Rmd”))
html	10d46c0	Jason Willwerscheid	2018-08-15	Build site.
Rmd	acd7f0c	Jason Willwerscheid	2018-08-15	wflow_publish(c(“analysis/index.Rmd”))
html	86ef31a	Jason Willwerscheid	2018-08-12	Build site.
Rmd	73aa339	Jason Willwerscheid	2018-08-12	wflow_publish(“analysis/index.Rmd”)
html	5a2e2a9	Jason Willwerscheid	2018-08-12	Build site.
Rmd	6a9bf1c	Jason Willwerscheid	2018-08-12	wflow_publish(c(“analysis/index.Rmd”, “analysis/parallel.Rmd”))
html	9246c68	Jason Willwerscheid	2018-08-12	Build site.
Rmd	eaba09b	Jason Willwerscheid	2018-08-12	wflow_publish(c(“analysis/index.Rmd”, “analysis/parallel.Rmd”))
html	cfa8e83	Jason Willwerscheid	2018-07-27	Build site.
Rmd	f497f36	Jason Willwerscheid	2018-07-27	wflow_publish(c(“analysis/warmstart2.Rmd”,
html	974e1db	Jason Willwerscheid	2018-07-27	Build site.
Rmd	2d81ef4	Jason Willwerscheid	2018-07-27	wflow_publish(c(“analysis/index.Rmd”, “analysis/init_fn2.Rmd”))
html	9fc59e7	Jason Willwerscheid	2018-07-26	Build site.
Rmd	21cbf89	Jason Willwerscheid	2018-07-26	wflow_publish(c(“analysis/index.Rmd”))
html	8f471cb	Jason Willwerscheid	2018-07-26	Build site.
Rmd	ff417b6	Jason Willwerscheid	2018-07-26	wflow_publish(c(“analysis/warmstart.Rmd”, “analysis/index.Rmd”))
html	aae4dfc	Jason Willwerscheid	2018-07-25	Build site.
Rmd	6417d4e	Jason Willwerscheid	2018-07-25	wflow_publish(c(“analysis/index.Rmd”, “analysis/init_fn.Rmd”))
html	3983695	Jason Willwerscheid	2018-07-25	Build site.
Rmd	566c202	Jason Willwerscheid	2018-07-25	wflow_publish(“analysis/index.Rmd”)
html	dd91b19	Jason Willwerscheid	2018-07-20	Build site.
Rmd	032edd6	Jason Willwerscheid	2018-07-20	wflow_publish(c(“analysis/index.Rmd”, “analysis/flash_em.Rmd”,
html	f995dbb	Jason Willwerscheid	2018-07-20	manual commits to remove licence
html	c9f10b0	Jason Willwerscheid	2018-07-20	Build site.
Rmd	4fc94bd	Jason Willwerscheid	2018-07-20	wflow_publish(c(“analysis/index.Rmd”, “analysis/flash_em.Rmd”))
html	1de2af0	Jason Willwerscheid	2018-07-20	Build site.
Rmd	9ac7bb0	Jason Willwerscheid	2018-07-20	wflow_publish(“analysis/index.Rmd”)
html	f98f61f	Jason Willwerscheid	2018-07-20	Build site.
Rmd	8e383cd	Jason Willwerscheid	2018-07-20	wflow_publish(c(“analysis/index.Rmd”))
html	58b3ea4	Jason Willwerscheid	2018-07-19	Build site.
Rmd	eb264b1	Jason Willwerscheid	2018-07-19	wflow_publish(“analysis/index.Rmd”)
html	36a6312	Jason Willwerscheid	2018-07-19	Build site.
Rmd	a9f8cc6	Jason Willwerscheid	2018-07-19	wflow_publish(“analysis/index.Rmd”)
html	c420802	Jason Willwerscheid	2018-07-17	Build site.
Rmd	38942eb	Jason Willwerscheid	2018-07-17	wflow_publish(“analysis/index.Rmd”)
html	e2e1286	Jason Willwerscheid	2018-07-16	Build site.
Rmd	55f42ca	Jason Willwerscheid	2018-07-16	wflow_publish(“analysis/index.Rmd”)
html	de8f823	Jason Willwerscheid	2018-07-16	Build site.
Rmd	f5379fa	Jason Willwerscheid	2018-07-16	wflow_publish(“analysis/index.Rmd”)
html	7db12a1	Jason Willwerscheid	2018-07-16	Build site.
html	873da40	Jason Willwerscheid	2018-07-16	Build site.
html	58e136a	Jason Willwerscheid	2018-07-15	Build site.
Rmd	82c7cdd	Jason Willwerscheid	2018-07-15	wflow_publish(c(“analysis/index.Rmd”, “analysis/about.Rmd”))
html	76eacb7	Jason Willwerscheid	2018-07-15	Build site.
Rmd	f6928cc	Jason Willwerscheid	2018-07-15	wflow_publish(c(“analysis/index.Rmd”, “analysis/about.Rmd”))
Rmd	56d7ded	Jason Willwerscheid	2018-07-14	Start workflowr project.

In Progress

Investigations 5a and 5b attempt to determine the best default initialization function.

Investigation 5a. We should change the default init_fn to udv_si_svd when there is missing data and udv_svd otherwise.
Investigation 5b. More evidence supporting a change to the default initialization function.

Investigations 8 and 9 implement parallel backfitting updates.

Investigation 8. Parallelizing the backfitting algorithm shows promise.
Investigation 9. An additional trick is needed to parallelize the backfitting updates performed in this MASH v FLASH GTEx analysis.

Investigation 10. SQUAREM does poorly on FLASH backfits. DAAREM (a more recent algorithm by one of the authors of SQUAREM) does better, but offers smaller performance gains than parallelization.

Investigation 11. The order in which factor/loading pairs are updated (during backfitting) makes some difference, but not much.

Still Relevant

Notes 1 and 2 and Investigation 4 describe a way to compute the FLASH objective directly (rather than using the indirect method implemented in flashr).

Note 1. Notes on computing the FLASH objective function. I derive an explicit expression for the KL divergence between prior and posterior.
Note 2. An alternate algorithm for optimizing the FLASH objective, using the explicit expression derived in the previous note.
Investigation 4. The alternate algorithm agrees with FLASH with respect to both the objective and fit obtained.

Archived

The bug causing the problem described in Investigations 1-3 was fixed in version 0.1-13 of package ebnm.

Investigation 1. The FLASH objective function can behave very erratically.
Investigation 2. This problem only occurs when using ebnm_pn, not ebnm_ash.
Investigation 3. The objective can continue to get worse as loadings are repeatedly updated. Nonetheless, convergence takes place (from above!).

Investigations 6 and 7 deal with warmstarts, which were implemented in version 0.5-14 of flashr.

Investigation 6. Poor optim results can produce large decreases in the objective function. We should use warmstarts when ebnm_fn = ebnm_pn.
Investigation 7. The advantages of warmstarts are not nearly as compelling when ebnm_fn = ebnm_ash.

This reproducible R Markdown analysis was created with workflowr 1.0.1