hOMEPAGe

workflowr

• Summary
• Checks
• Past versions

Last updated: 2023-09-12

Checks: 2 0

Knit directory: gsmash/

This reproducible R Markdown analysis was created with workflowr (version 1.6.2). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.

---

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: 3559787

Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.

The results in this page were generated with repository version 3559787. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.

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:    .Rhistory
    Ignored:    .Rproj.user/

Untracked files:
    Untracked:  analysis/GO_ORA_montoro.Rmd
    Untracked:  analysis/GO_ORA_pbmc_purified.Rmd
    Untracked:  analysis/fit_ebpmf_sla_2000.Rmd
    Untracked:  chipexo_rep1_reverse.rds
    Untracked:  data/Citation.RData
    Untracked:  data/SLA/
    Untracked:  data/abstract.txt
    Untracked:  data/abstract.vocab.txt
    Untracked:  data/ap.txt
    Untracked:  data/ap.vocab.txt
    Untracked:  data/sla_2000.rds
    Untracked:  data/sla_full.rds
    Untracked:  data/text.R
    Untracked:  data/tpm3.rds
    Untracked:  output/driving_gene_pbmc.rds
    Untracked:  output/pbmc_gsea.rds
    Untracked:  output/plots/
    Untracked:  output/tpm3_fit_fasttopics.rds
    Untracked:  output/tpm3_fit_stm.rds
    Untracked:  output/tpm3_fit_stm_slow.rds
    Untracked:  sla.rds

Unstaged changes:
    Modified:   analysis/PMF_splitting.Rmd
    Modified:   analysis/fit_ebpmf_sla.Rmd
    Modified:   code/poisson_STM/structure_plot.R
    Modified:   code/poisson_mean/pois_log_normal_mle.R

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.

---

These are the previous versions of the repository in which changes were made to the R Markdown (analysis/index.Rmd) and HTML (docs/index.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.

File	Version	Author	Date	Message
Rmd	3559787	DongyueXie	2023-09-12	wflow_publish("analysis/index.Rmd")
html	e568806	DongyueXie	2023-08-07	Build site.
Rmd	d0c8224	DongyueXie	2023-08-07	wflow_publish(c("analysis/fit_ebpmf_sla_nonneg.Rmd", "analysis/fit_ebpmf_sla_full_nonneg.Rmd",
html	5b924fb	DongyueXie	2023-07-06	Build site.
Rmd	1a93f11	DongyueXie	2023-07-06	wflow_publish("analysis/index.Rmd")
Rmd	ad9ee73	DongyueXie	2023-04-22	update
html	f7c1f70	DongyueXie	2023-03-31	Build site.
Rmd	fcaf7de	DongyueXie	2023-03-31	wflow_publish("analysis/index.Rmd")
html	9bce1b1	DongyueXie	2023-03-05	Build site.
Rmd	ed26457	DongyueXie	2023-03-05	wflow_publish("analysis/index.Rmd")
html	8db89ac	DongyueXie	2023-02-27	Build site.
Rmd	55b0c40	DongyueXie	2023-02-27	wflow_publish("analysis/index.Rmd")
html	5ca9972	DongyueXie	2023-02-26	Build site.
Rmd	dd08b27	DongyueXie	2023-02-26	wflow_publish("analysis/index.Rmd")
html	c49a513	DongyueXie	2023-02-24	Build site.
Rmd	e97912c	DongyueXie	2023-02-24	wflow_publish("analysis/index.Rmd")
html	a500ed1	DongyueXie	2023-02-23	Build site.
Rmd	5eb6fd5	DongyueXie	2023-02-23	wflow_publish("analysis/index.Rmd")
html	8546abc	DongyueXie	2023-02-23	Build site.
Rmd	75d321e	DongyueXie	2023-02-23	wflow_publish("analysis/index.Rmd")
html	5366917	DongyueXie	2023-02-22	Build site.
Rmd	83b569b	DongyueXie	2023-02-22	wflow_publish("analysis/index.Rmd")
html	d5da1a6	DongyueXie	2023-02-21	Build site.
Rmd	36dabc4	DongyueXie	2023-02-21	wflow_publish("analysis/index.Rmd")
html	16c909c	DongyueXie	2023-02-20	Build site.
Rmd	5f313a2	DongyueXie	2023-02-20	wflow_publish("analysis/index.Rmd")
html	0503ead	DongyueXie	2023-02-20	Build site.
Rmd	9c68f5d	DongyueXie	2023-02-20	wflow_publish("analysis/index.Rmd")
html	f53734a	DongyueXie	2023-02-19	Build site.
Rmd	0c85a75	DongyueXie	2023-02-19	wflow_publish("analysis/index.Rmd")
html	0ff8393	DongyueXie	2023-02-09	Build site.
Rmd	b067364	DongyueXie	2023-02-09	wflow_publish("analysis/index.Rmd")
html	0915d9a	DongyueXie	2023-02-07	Build site.
Rmd	1019323	DongyueXie	2023-02-07	wflow_publish("analysis/index.Rmd")
html	a6cff57	DongyueXie	2023-01-27	Build site.
Rmd	60d225b	DongyueXie	2023-01-27	wflow_publish("analysis/index.Rmd")
html	4d7d3c8	DongyueXie	2023-01-27	Build site.
Rmd	3210f97	DongyueXie	2023-01-27	wflow_publish("analysis/index.Rmd")
Rmd	6ab6622	DongyueXie	2023-01-26	server donw
html	2bcf9f6	DongyueXie	2023-01-11	Build site.
Rmd	db0dc5c	DongyueXie	2023-01-11	wflow_publish("analysis/index.Rmd")
html	4679d22	DongyueXie	2023-01-09	Build site.
Rmd	9648eea	DongyueXie	2023-01-09	wflow_publish("analysis/index.Rmd")
html	0e2cdec	DongyueXie	2023-01-07	Build site.
Rmd	44b820b	DongyueXie	2023-01-07	wflow_publish("analysis/index.Rmd")
html	e97f793	DongyueXie	2023-01-05	Build site.
Rmd	4247caa	DongyueXie	2023-01-05	wflow_publish("analysis/index.Rmd")
html	dda8cc9	DongyueXie	2023-01-05	Build site.
Rmd	7bb2272	DongyueXie	2023-01-05	wflow_publish("analysis/index.Rmd")
html	f1d89f4	DongyueXie	2023-01-04	Build site.
Rmd	08ef090	DongyueXie	2023-01-04	wflow_publish("analysis/index.Rmd")
html	3c121ef	DongyueXie	2023-01-02	Build site.
Rmd	13800ee	DongyueXie	2023-01-02	wflow_publish("analysis/index.Rmd")
html	d6b94d9	DongyueXie	2023-01-01	Build site.
Rmd	d4437e6	DongyueXie	2023-01-01	wflow_publish("analysis/index.Rmd")
html	1109340	DongyueXie	2022-12-29	Build site.
Rmd	5bb04de	DongyueXie	2022-12-29	wflow_publish("analysis/index.Rmd")
html	15e17ec	DongyueXie	2022-12-27	Build site.
Rmd	22b3cb3	DongyueXie	2022-12-27	wflow_publish("analysis/index.Rmd")
html	0ce12e6	DongyueXie	2022-12-25	Build site.
Rmd	3e0c642	DongyueXie	2022-12-25	wflow_publish("analysis/index.Rmd")
html	0781406	DongyueXie	2022-12-19	Build site.
Rmd	0b57096	DongyueXie	2022-12-19	wflow_publish("analysis/index.Rmd")
html	11d7ce7	DongyueXie	2022-12-19	Build site.
Rmd	e0212d0	DongyueXie	2022-12-19	wflow_publish("analysis/index.Rmd")
html	e0212d0	DongyueXie	2022-12-19	wflow_publish("analysis/index.Rmd")
html	2368e4d	DongyueXie	2022-12-19	Build site.
Rmd	7365c10	DongyueXie	2022-12-19	wflow_publish("analysis/index.Rmd")
html	1cf2766	DongyueXie	2022-12-18	Build site.
Rmd	a870020	DongyueXie	2022-12-18	wflow_publish("analysis/index.Rmd")
html	c409906	DongyueXie	2022-12-17	Build site.
Rmd	3481595	DongyueXie	2022-12-17	wflow_publish("analysis/index.Rmd")
html	d94d5c1	DongyueXie	2022-12-17	Build site.
Rmd	3abf967	DongyueXie	2022-12-17	wflow_publish("analysis/index.Rmd")
html	9e9974e	Dongyue Xie	2022-12-16	Build site.
Rmd	4ec37a2	Dongyue Xie	2022-12-16	wflow_publish("analysis/index.Rmd")
html	dcd5d99	DongyueXie	2022-12-09	Build site.
Rmd	6c9f108	DongyueXie	2022-12-09	wflow_publish("analysis/index.Rmd")
html	fce6a08	DongyueXie	2022-12-07	Build site.
Rmd	02435cd	DongyueXie	2022-12-07	wflow_publish("analysis/index.Rmd")
html	f626fbd	DongyueXie	2022-12-06	Build site.
Rmd	ddd4418	DongyueXie	2022-12-06	wflow_publish("analysis/index.Rmd")
html	ba6d36b	DongyueXie	2022-12-06	Build site.
Rmd	569eba8	DongyueXie	2022-12-06	wflow_publish("analysis/index.Rmd")
html	d2d140f	DongyueXie	2022-12-04	Build site.
Rmd	65c5c38	DongyueXie	2022-12-04	wflow_publish("analysis/index.Rmd")
html	035e735	DongyueXie	2022-11-28	Build site.
Rmd	a3c7d8a	DongyueXie	2022-11-28	wflow_publish("analysis/index.Rmd")
html	8aed873	DongyueXie	2022-11-28	Build site.
Rmd	5e3d5b1	DongyueXie	2022-11-28	wflow_publish("analysis/index.Rmd")
html	7953ed0	DongyueXie	2022-11-22	Build site.
Rmd	343de7b	DongyueXie	2022-11-22	wflow_publish("analysis/index.Rmd")
html	b855ed1	DongyueXie	2022-11-21	Build site.
Rmd	6cd3705	DongyueXie	2022-11-21	wflow_publish("analysis/index.Rmd")
html	c773488	DongyueXie	2022-11-18	Build site.
Rmd	c6ad06f	DongyueXie	2022-11-18	wflow_publish("analysis/index.Rmd")
html	a7150cf	DongyueXie	2022-11-17	Build site.
Rmd	5e4038f	DongyueXie	2022-11-17	wflow_publish("analysis/index.Rmd")
html	82b7d06	DongyueXie	2022-11-17	Build site.
Rmd	1642707	DongyueXie	2022-11-17	wflow_publish("analysis/index.Rmd")
html	97ed5ae	DongyueXie	2022-11-16	Build site.
Rmd	8428e51	DongyueXie	2022-11-16	wflow_publish("analysis/index.Rmd")
html	befb8ea	DongyueXie	2022-11-15	Build site.
Rmd	a6aa06c	DongyueXie	2022-11-15	wflow_publish("analysis/index.Rmd")
html	3f4cd85	DongyueXie	2022-11-15	Build site.
Rmd	d45c20f	DongyueXie	2022-11-15	wflow_publish("analysis/index.Rmd")
html	dd4d101	DongyueXie	2022-11-15	Build site.
Rmd	31ff30d	DongyueXie	2022-11-15	wflow_publish("analysis/index.Rmd")
html	ef38082	DongyueXie	2022-11-15	Build site.
Rmd	340c808	DongyueXie	2022-11-15	wflow_publish("analysis/index.Rmd")
html	79bc4c4	DongyueXie	2022-11-12	Build site.
Rmd	902ec36	DongyueXie	2022-11-12	wflow_publish("analysis/index.Rmd")
html	aca3b1f	DongyueXie	2022-11-11	Build site.
Rmd	04ec5b3	DongyueXie	2022-11-11	wflow_publish("analysis/index.Rmd")
html	86f98d5	DongyueXie	2022-10-28	Build site.
Rmd	9f29c63	DongyueXie	2022-10-28	wflow_publish("analysis/index.Rmd")
html	e283ad6	DongyueXie	2022-10-14	Build site.
Rmd	8694b47	DongyueXie	2022-10-14	wflow_publish("analysis/index.Rmd")
html	65ba198	DongyueXie	2022-10-14	Build site.
Rmd	cf17f42	DongyueXie	2022-10-14	wflow_publish("analysis/index.Rmd")
html	c8ee67d	DongyueXie	2022-10-12	Build site.
Rmd	8d449b7	DongyueXie	2022-10-12	wflow_publish("analysis/index.Rmd")
html	49f65d9	DongyueXie	2022-10-12	Build site.
Rmd	0449083	DongyueXie	2022-10-12	wflow_publish("analysis/index.Rmd")
html	d82d960	Dongyue Xie	2022-10-11	Build site.
Rmd	41da454	Dongyue Xie	2022-10-11	wflow_publish("analysis/index.Rmd")
html	9263a71	Dongyue Xie	2022-10-06	Build site.
Rmd	75bcfb3	Dongyue Xie	2022-10-06	wflow_publish("analysis/index.Rmd")
html	c3a4a4d	Dongyue Xie	2022-10-03	Build site.
Rmd	b73cc63	Dongyue Xie	2022-10-03	wflow_publish("analysis/index.Rmd")
html	4e18aa6	Dongyue Xie	2022-09-29	Build site.
Rmd	967deea	Dongyue Xie	2022-09-29	add inversion method
html	c8343ef	Dongyue Xie	2022-09-27	Build site.
Rmd	1e05e86	Dongyue Xie	2022-09-27	wflow_publish("analysis/index.Rmd")
html	5fb1cad	Dongyue Xie	2022-06-06	Build site.
Rmd	8b15558	Dongyue Xie	2022-06-06	Publish the initial files for myproject
html	a3f39a4	Dongyue Xie	2022-06-06	Build site.
Rmd	e87f570	Dongyue Xie	2022-06-06	none
html	76d4095	Dongyue Xie	2022-06-06	Build site.
Rmd	9df5fb2	Dongyue Xie	2022-06-06	Start my new project
Rmd	c28e1a6	Dongyue Xie	2022-06-06	Start workflowr project.

---

Poisson Matrix Factorization

Analysis

• Negative Binomial and Poisson log-normal mean-var relationship
• mean variance relationship in real single cell data
• attempt to reduce run time
• memory usage analysis
• size of sigma2 on convergence speed
• fit K=1 on pbmc data
• [point exponential and unimodal nonnegative]
• ebpm normal prior init in PMF
• ebpmf vs topic model
• Poisson deviance
• Multiplicative vs additive
• smallman example

Simulation

• Splitting method in PMF, simple simulation example
• Simulation based on PBMC data, 3 cell types
• Simulation based on PBMC data, 3 cell types, 1 init backfitting
• Simulation pbmc data from fasttopics, nonnegtive L and F
• Simulation data from fitted model on pbmc data in fasttopics, nonnegtive L and F
• Simulation data from fitted model on pbmc data in fasttopics, nonnegtive L and F, added more models
• Simulation data from fitted model on pbmc data in fasttopics, nonnegtive L and F, true sigma2 = 0

On text data

SLA data set

• EBPMF does not find factors?
• Pre-processing SLA data
• EBPMF does not find factors?
• log transformation + EBNMF on SLA data
• ebpmf init using liklihood expansion
• methods fit on sla 2000 data
• methods fit on sla full data, filter out words in less than 5 doc
• methods fit on sla full data, filter out words in less than 3 doc

On pbmc3k

• pbmc data in fasttopics package, flash, glmpca, and topic model fit
• pbmc3k data 10X, flash, glmpca, and topic model fit
• pbmc3k 10X result
• pbmc in fasttopic results
• pbmc in fasttopic,effect of initialization variances
• pbmc in fasttopic, 3-5 cell types for simulation

Below are some analysis I have done, before putting \(l_0,f_0\) as part of flash, and estimating \(f_0\).

• run PMF on PBMC data, 3 cell types
• run PMF on PBMC data, 3 cell types, nonnegative loadings
• run PMF on PBMC data, 3 cell types, nonnegative loadings and factors
• Compare VGA iterations on PBMC data, 3 cell types
• Compare VGA iterations on PBMC data, full dataset
• run PMF on PBMC data
• run PMF on PBMC data,nonnegative Loadings
• run PMF on PBMC data,nonnegative Loadings and factors
• library size distribution
• gene specific variance and gene expression
• Compare latent M from splitting and glmpca
• Run flash on latent M from glmpca
• How do results differ over iterations? When can we stop iterations?
• Use GLMPCA as init
• estimate f0, nonnegative LF case

On Droplet

• droplet Montoro results
• droplet Montoro results, no filtering of genes

On PBMC purified

• PBMC purified result

On PBMC 68K

• PBMC 68K result

On Pancreas data

Poisson mean

• Poisson log normal
• NB JJ lower bound
• NB PG augmentation
• NB approxiamation to Poisson
• Poisson mean \(\log(1+\exp(\cdot))\) link
• VGA objective function v(m) relationship
• Poisson mean splitting local optimum

Penalized

• Poisson mean penalized form, compound
• Poisson mean penalized form, inversion

Benchmarking

• log link
• exp prior
• gamma prior
• t prior

With Run time Comparison:

• exp prior with runtime
• log prior with runtime

Overdispersed model

• generate data from splitting model
• generate data from negative binomial model
• thoughts on the results

Poisson smoothing

• The splitting method smoothing overdispersed count data
• Poisson smooth splitting local optimum

Real data

• Smooth RNA seq

Bechmarking smoothing methods

Some old results

Details

• low count smooth ndwt

Binomial

• Apply Splitting to Binomial Mean problem
• It does not work for binary data
• Apply Splitting to Binomial smooth problem
• is vga binom problem convex?

Normal mean penalty form

• bound on (log g)’’
• log density of normal mixture
  • log density of normal mixture, variance ratio
• log density of laplace mixture
• normal mean penalty plots
• Solve normal mean problem using optimization - compund and inversion penalty
• Solve normal mean problem using optimization - also estimate prior mean
• normal mean penalty plots, when \(s^2\) depends on posterior mean
• normal mean penalty plots, when \(s^2\) depends on posterior mean, penalty exlcuding \(\log(2\pi s^2)/2\)
• where is the log term from in the penalty