Meta-analysis with MMUPHin and Multi-omics Integration with Cooperative Learning #394

YihanLiu4023 · 2024-02-28T16:25:26Z

New PR for 2 new chapters: meta-analysis with MMUPHin package and multi-omics integration with cooperative learning.

TuomasBorman

Hello!

First of all thank you for this PR. Looks very nice!! Sorry for taking this long to respond. Unfortunately, I haven't had time to review this in detail. However with quick check, I noticed some things that you might want to consider. I will get back to you when I have checked this more thoroughly.

-Tuomas

inst/pages/IntegratedLearner.qmd

TuomasBorman

All the packages that are used, should be in DESCRIPTION file
Some chunk options were incorrect

I think with these changes this is good to go

inst/pages/IntegratedLearner.qmd

inst/pages/MSEA.qmd

inst/pages/IntegratedLearner.qmd

inst/pages/MMUPHin_meta_analysis.qmd

inst/pages/IntegratedLearner.qmd

inst/pages/MMUPHin_meta_analysis.qmd

antagomir · 2024-04-10T20:25:41Z

Please resolve the review discussions when they have been handled. This way we can keep track on possible issues here that are still pending. When no discussion points are open it is a good sign that this is getting ready.

antagomir · 2024-04-17T22:05:17Z

Only very small final changes needed, it seems to me.

@YihanLiu4023 - let us know if you need any support from our side in finalizing these. This is a good opportunity to get in sync with the general development guidelines.

antagomir · 2024-04-20T20:03:39Z

Hi! I see see that you're working on this @YihanLiu4023 - just confirm when ready, so we know to proceed checking.

YihanLiu4023 · 2024-04-22T03:23:59Z

Hi! I see see that you're working on this @YihanLiu4023 - just confirm when ready, so we know to proceed checking.

Sure! Thank you for the reminder, I have already uploaded all the updates. Please feel free to proceed the checking.

TuomasBorman

I think this is now good to go. Some changes related to build still needed but I can do them.

These chapters are very nice! Thanks!

inst/pages/MMUPHin_meta_analysis.qmd

antagomir · 2024-04-22T08:19:14Z

inst/pages/MMUPHin_meta_analysis.qmd

+
+library(dplyr)
+
+se_relative <- sampleMetadata |>


It seems to me that this is DataFrame (metadata) rather than SummarizedExperiment object (abbreviation "se").

If so, I suggest to rename "se_relative" and "se_pathway" in order to avoid confusing them with SE objects?

Good point. Thanks for the comment!

antagomir · 2024-04-22T08:19:28Z

inst/pages/MMUPHin_meta_analysis.qmd

+
+```{r load_curatedMetagenomicData}
+library(curatedMetagenomicData)
+


remove the empty line 66?

antagomir · 2024-04-22T09:17:23Z

inst/pages/MMUPHin_meta_analysis.qmd

+##############
+# Before
+R2_before <- round(fit_adonis_before$aov.tab[1, 5]*100, 1)
+pcoa_before <- cmdscale(D_before, eig = TRUE)
+ord_before <- as.data.frame(pcoa_before$points)
+percent_var_before <- round(pcoa_before$eig / sum(pcoa_before$eig) * 100, 1)[1:2]
+before_labels <- c(paste('Axis 1 (', percent_var_before[1], '%)', sep = ''), 
+                 paste('Axis 2 (', percent_var_before[2], '%)', sep = ''))
+
+before_phrase <- paste('Before (PERMANOVA R2 = ', R2_before, '%)', sep = '')
+colnames(ord_before) <- c('PC1', 'PC2')
+ord_before$Study <- data_meta$study_name
+
+# After
+R2_after <- round(fit_adonis_after$aov.tab[1, 5]*100, 1)
+pcoa_after <- cmdscale(D_after, eig = TRUE)
+ord_after <- as.data.frame(pcoa_after$points)
+percent_var_after <- round(pcoa_after$eig / sum(pcoa_after$eig) * 100, 1)[1:2]
+after_labels <- c(paste('Axis 1 (', percent_var_after[1], '%)', sep = ''), 
+                paste('Axis 2 (', percent_var_after[2], '%)', sep = ''))
+
+after_phrase <- paste('After (PERMANOVA R2 = ', R2_after, '%)', sep = '')
+colnames(ord_after) <- c('PC1', 'PC2')
+ord_after$Study <- data_meta$study_name


same, mia/SE tools would be rather used (section 7 in OMA has examples)

antagomir · 2024-04-22T09:17:46Z

inst/pages/MMUPHin_meta_analysis.qmd

+# Ordination Plot
+p_before <- ggplot(ord_before, aes(x = PC1, y = PC2, color = Study)) + 
+  geom_point(size = 4) + 
+  theme_bw() + 
+  xlab(before_labels[1]) + 
+  ylab(before_labels[2]) + 
+  ggtitle(before_phrase) + 
+  theme(plot.title = element_text(hjust = 0.5)) + 
+  theme(legend.position ="none")
+
+p_after <- ggplot(ord_after, aes(x = PC1, y = PC2, color = Study)) + 
+  geom_point(size = 4) + 
+  theme_bw() + 
+  xlab(after_labels[1]) + 
+  ylab(after_labels[2]) + 
+  ggtitle(after_phrase) + 
+  theme(plot.title = element_text(hjust = 0.5)) + 
+  theme(legend.position ="none")
+


We have dedicated methods for ordination plots in mia / miaViz

antagomir · 2024-04-22T09:18:32Z

inst/pages/MMUPHin_meta_analysis.qmd

+```{r prepare-input-data-pathway-abundance}
+


Here also, could we prepare data in SE format, then use tools dedicated to SE.

antagomir · 2024-04-22T09:18:46Z

inst/pages/MMUPHin_meta_analysis.qmd

+D_before <- vegdist(t(data_abd))
+D_after <- vegdist(t(data_abd_adj))
+
+set.seed(1)
+fit_adonis_before <- adonis(D_before ~ study_name, data = data_meta)
+fit_adonis_after <- adonis(D_after ~ study_name, data = data_meta)


antagomir · 2024-04-22T09:18:59Z

inst/pages/MMUPHin_meta_analysis.qmd

+# Before
+R2_before <- round(fit_adonis_before$aov.tab[1, 5]*100, 1)
+pcoa_before <- cmdscale(D_before, eig = TRUE)
+ord_before <- as.data.frame(pcoa_before$points)
+percent_var_before <- round(pcoa_before$eig / sum(pcoa_before$eig) * 100, 1)[1:2]
+before_labels <- c(paste('Axis 1 (', percent_var_before[1], '%)', sep = ''), 
+                 paste('Axis 2 (', percent_var_before[2], '%)', sep = ''))
+
+before_phrase <- paste('Before (PERMANOVA R2 = ', R2_before, '%)', sep = '')
+colnames(ord_before) <- c('PC1', 'PC2')
+ord_before$Study <- data_meta$study_name
+
+# After
+R2_after <- round(fit_adonis_after$aov.tab[1, 5]*100, 1)
+pcoa_after <- cmdscale(D_after, eig = TRUE)
+ord_after <- as.data.frame(pcoa_after$points)


antagomir · 2024-04-22T09:24:28Z

Ok thanks. Now with the more detailed look I noticed one major issue, which is that the scripts are primarily using base R & vegan, rather than the TreeSummarizedExperiment framework. This is somewhat problematic since the entire OMA gitbook is built on the idea of supporting the standardization of methods and shift towards the TreeSummarizedExperiment ecosystem of tools.

Now I think we have two options:

Update the current PR to use SE framework, before merging
Merge now, and then update afterwards, ideally by the original contributor @YihanLiu4023 and/or with support from the broader contributor base. This latter option has the benefit that we can proceed with other things that are pending on merging this PR but this also has a risk that these necessary updates towards SE ecosystem will be delayed.

@YihanLiu4023 @TuomasBorman your comments will be appreciated

antagomir · 2024-04-22T09:25:34Z

Ok I can see that this was merged already so I guess we are going with the option 2.

@YihanLiu4023 would you be able to make some of the suggested updates (above pending comments) in a new PR?

antagomir · 2024-04-22T10:42:47Z

If the idea is to update this to IntegratedLearning and that will support SE/MAE, then we can deal with these in the next round?

TuomasBorman · 2024-04-22T11:01:12Z

I agree with your points @antagomir . Sorry for merging it before you had time to give your review.

The material in chapters are very good addition but now they are not using TreeSE ecosystem which is a big lack. However, I believe it was better to merge now and improve later. I understood that these chapters will be updated later. (Also great chance to integrate TreeSE to methods that are not supporting it yet)

antagomir · 2024-04-22T11:09:53Z

@YihanLiu4023 could you summarize the planned updates and their planned schedule?

antagomir · 2024-04-22T17:55:09Z

I just noticed that these new chapters are not rendered as part of the OMA gitbook https://microbiome.github.io/OMA/

I think the reason is that the qmd file names have not been listed here:
inst/assets/_book.yml

@YihanLiu4023 did you check that you can render and view the book locally, before opening the PR?

TuomasBorman · 2024-04-22T18:07:56Z

This issue is already handled with PR https://github.com/microbiome/OMA/pull/433/files

himelmallick · 2024-04-23T13:56:57Z

I agree that going with Option 2 and updating later. We will make sure to use the SE/TSE framework in the IntegratedLearner R package which can be added back to the chapters instead of the base R functions and the same thing applies to the other packages. Thanks all for your comments. @YihanLiu4023 will work on the immediately addressable comments from above.

antagomir · 2024-04-23T16:30:33Z

great to hear, we needed to merge but let's continue and take it forward together; don't hesitate to ask if anything is unclear as the ecosystem and documentation are shaping up and we can improve where necessary

YihanLiu4023 added 3 commits February 28, 2024 11:11

Add files via upload

04fd860

Add files via upload

d58b360

Add files via upload

11552ec

TuomasBorman requested changes Mar 10, 2024

View reviewed changes

inst/pages/IntegratedLearner.qmd Outdated Show resolved Hide resolved

inst/pages/IntegratedLearner.qmd Outdated Show resolved Hide resolved

inst/pages/IntegratedLearner.qmd Outdated Show resolved Hide resolved

inst/pages/IntegratedLearner.qmd Outdated Show resolved Hide resolved

YihanLiu4023 added 3 commits March 21, 2024 11:01

Add files via upload

6e24f74

Add files via upload

38448eb

Delete inst/GSEA.qmd

fa73146

TuomasBorman reviewed Mar 21, 2024

View reviewed changes

inst/pages/IntegratedLearner.qmd Outdated Show resolved Hide resolved

YihanLiu4023 added 7 commits March 26, 2024 05:37

Update IntegratedLearner.qmd

7b4d5d6

Update MMUPHin_meta_analysis.qmd

93252df

Update DESCRIPTION

be305d0

Update MMUPHin_meta_analysis.qmd

c485f98

Update IntegratedLearner.qmd

4c2a41f

Delete inst/pages/images/GSEA.qmd

972c40d

Add files via upload

69a7365

TuomasBorman requested changes Apr 9, 2024

View reviewed changes

Update 90_acknowledgments.qmd

56acf65

YihanLiu4023 added 3 commits April 15, 2024 04:31

Update DESCRIPTION

55d985a

Update bibliography.bib

844dd32

Update bibliography.bib

235fe8d

antagomir and others added 5 commits April 18, 2024 01:08

Merge branch 'master' into master

1b50ef4

Update MMUPHin_meta_analysis.qmd

c6d0926

Update MSEA.qmd

a9dde8f

Update MMUPHin_meta_analysis.qmd

1e1006f

Update IntegratedLearner.qmd

88a637a

TuomasBorman approved these changes Apr 22, 2024

View reviewed changes

TuomasBorman merged commit d3264df into microbiome:master Apr 22, 2024

antagomir requested changes Apr 22, 2024

View reviewed changes

antagomir mentioned this pull request Apr 22, 2024

Integrate new multiomic chapters in the book view #437

Open

Daenarys8 mentioned this pull request May 6, 2024

Update Multi-omics chapter to use SE/TSE framework #477

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Meta-analysis with MMUPHin and Multi-omics Integration with Cooperative Learning #394

Meta-analysis with MMUPHin and Multi-omics Integration with Cooperative Learning #394

YihanLiu4023 commented Feb 28, 2024

TuomasBorman left a comment

TuomasBorman left a comment

antagomir commented Apr 10, 2024

antagomir commented Apr 17, 2024

antagomir commented Apr 20, 2024

YihanLiu4023 commented Apr 22, 2024

TuomasBorman left a comment

antagomir Apr 22, 2024

himelmallick Apr 23, 2024

antagomir Apr 22, 2024

antagomir Apr 22, 2024

antagomir Apr 22, 2024

antagomir Apr 22, 2024

antagomir Apr 22, 2024

antagomir Apr 22, 2024

antagomir commented Apr 22, 2024

antagomir commented Apr 22, 2024

antagomir commented Apr 22, 2024

TuomasBorman commented Apr 22, 2024

antagomir commented Apr 22, 2024

antagomir commented Apr 22, 2024

TuomasBorman commented Apr 22, 2024

himelmallick commented Apr 23, 2024

antagomir commented Apr 23, 2024


		```{r load_curatedMetagenomicData}
		library(curatedMetagenomicData)


		library(dplyr)

		se_relative <- sampleMetadata \|>

Meta-analysis with MMUPHin and Multi-omics Integration with Cooperative Learning #394

Meta-analysis with MMUPHin and Multi-omics Integration with Cooperative Learning #394

Conversation

YihanLiu4023 commented Feb 28, 2024

TuomasBorman left a comment

Choose a reason for hiding this comment

TuomasBorman left a comment

Choose a reason for hiding this comment

antagomir commented Apr 10, 2024

antagomir commented Apr 17, 2024

antagomir commented Apr 20, 2024

YihanLiu4023 commented Apr 22, 2024

TuomasBorman left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

antagomir commented Apr 22, 2024

antagomir commented Apr 22, 2024

antagomir commented Apr 22, 2024

TuomasBorman commented Apr 22, 2024

antagomir commented Apr 22, 2024

antagomir commented Apr 22, 2024

TuomasBorman commented Apr 22, 2024

himelmallick commented Apr 23, 2024

antagomir commented Apr 23, 2024