Samuel Marsh Ph.D.
@samuelmarsh.bsky.social
Postdoc Beth Stevens Lab | Interested in Alzheimer’s disease, all things single cell, and neuroimmune | Dad 👧👶🏻 | #scicomm #rstats | NEU ‘10 & UCI ’16
https://samuel-marsh.science
https://samuel-marsh.science
Come check out our whole session (starts at 4:15pm) with excellent talks from collaborators Evan Macosko, Henrik Zetterberg, Ville Leinonen, & Tarja Malm.
𝙎𝙩𝙪𝙙𝙮𝙞𝙣𝙜 𝙩𝙞𝙨𝙨𝙪𝙚 𝙨𝙖𝙢𝙥𝙡𝙚𝙨 𝙛𝙧𝙤𝙢 𝙧𝙖𝙧𝙚 𝙘𝙤𝙝𝙤𝙧𝙩𝙨 𝙩𝙤 𝙡𝙚𝙖𝙧𝙣 𝙖𝙗𝙤𝙪𝙩 𝙚𝙖𝙧𝙡𝙮 𝙨𝙩𝙖𝙜𝙚 𝘼𝘿 𝙥𝙖𝙩𝙝𝙤𝙥𝙝𝙮𝙨𝙞𝙤𝙡𝙤𝙜𝙮.
𝙎𝙩𝙪𝙙𝙮𝙞𝙣𝙜 𝙩𝙞𝙨𝙨𝙪𝙚 𝙨𝙖𝙢𝙥𝙡𝙚𝙨 𝙛𝙧𝙤𝙢 𝙧𝙖𝙧𝙚 𝙘𝙤𝙝𝙤𝙧𝙩𝙨 𝙩𝙤 𝙡𝙚𝙖𝙧𝙣 𝙖𝙗𝙤𝙪𝙩 𝙚𝙖𝙧𝙡𝙮 𝙨𝙩𝙖𝙜𝙚 𝘼𝘿 𝙥𝙖𝙩𝙝𝙤𝙥𝙝𝙮𝙨𝙞𝙤𝙡𝙤𝙜𝙮.
July 27, 2025 at 1:42 PM
Come check out our whole session (starts at 4:15pm) with excellent talks from collaborators Evan Macosko, Henrik Zetterberg, Ville Leinonen, & Tarja Malm.
𝙎𝙩𝙪𝙙𝙮𝙞𝙣𝙜 𝙩𝙞𝙨𝙨𝙪𝙚 𝙨𝙖𝙢𝙥𝙡𝙚𝙨 𝙛𝙧𝙤𝙢 𝙧𝙖𝙧𝙚 𝙘𝙤𝙝𝙤𝙧𝙩𝙨 𝙩𝙤 𝙡𝙚𝙖𝙧𝙣 𝙖𝙗𝙤𝙪𝙩 𝙚𝙖𝙧𝙡𝙮 𝙨𝙩𝙖𝙜𝙚 𝘼𝘿 𝙥𝙖𝙩𝙝𝙤𝙥𝙝𝙮𝙨𝙞𝙤𝙡𝙤𝙜𝙮.
𝙎𝙩𝙪𝙙𝙮𝙞𝙣𝙜 𝙩𝙞𝙨𝙨𝙪𝙚 𝙨𝙖𝙢𝙥𝙡𝙚𝙨 𝙛𝙧𝙤𝙢 𝙧𝙖𝙧𝙚 𝙘𝙤𝙝𝙤𝙧𝙩𝙨 𝙩𝙤 𝙡𝙚𝙖𝙧𝙣 𝙖𝙗𝙤𝙪𝙩 𝙚𝙖𝙧𝙡𝙮 𝙨𝙩𝙖𝙜𝙚 𝘼𝘿 𝙥𝙖𝙩𝙝𝙤𝙥𝙝𝙮𝙨𝙞𝙤𝙡𝙤𝙜𝙮.
Reposted by Samuel Marsh Ph.D.
Implementation and validation of single-cell genomics experiments in neuroscience
www.nature.com/articles/s41...
Validation is key! This review dives into best practices for validating ssc/snRNAseq experiments. Learn about the challenges and solutions for ensuring robust and reliable findings.
www.nature.com/articles/s41...
Validation is key! This review dives into best practices for validating ssc/snRNAseq experiments. Learn about the challenges and solutions for ensuring robust and reliable findings.
a cartoon character is holding a can of beans with the words how do i use it below him
ALT: a cartoon character is holding a can of beans with the words how do i use it below him
media.tenor.com
December 11, 2024 at 5:04 PM
Implementation and validation of single-cell genomics experiments in neuroscience
www.nature.com/articles/s41...
Validation is key! This review dives into best practices for validating ssc/snRNAseq experiments. Learn about the challenges and solutions for ensuring robust and reliable findings.
www.nature.com/articles/s41...
Validation is key! This review dives into best practices for validating ssc/snRNAseq experiments. Learn about the challenges and solutions for ensuring robust and reliable findings.
Finally, as always, I want to thank everyone who uses scCustomize. I could never have imagined it would be adopted and used by some many people and it’s really fantastic every time I read paper or see plot that was made using scCustomize.
Happy single cell analysis!!! 💻📈🧬
20/20
Happy single cell analysis!!! 💻📈🧬
20/20
December 6, 2024 at 4:36 PM
Finally, as always, I want to thank everyone who uses scCustomize. I could never have imagined it would be adopted and used by some many people and it’s really fantastic every time I read paper or see plot that was made using scCustomize.
Happy single cell analysis!!! 💻📈🧬
20/20
Happy single cell analysis!!! 💻📈🧬
20/20
As always there are a ton of bug fixes in this update here too. Thank you to everyone who has report issues on GitHub!
There’s more in this update too besides these highlights so check out changelog for full details!!
19/n
There’s more in this update too besides these highlights so check out changelog for full details!!
19/n
December 6, 2024 at 4:36 PM
As always there are a ton of bug fixes in this update here too. Thank you to everyone who has report issues on GitHub!
There’s more in this update too besides these highlights so check out changelog for full details!!
19/n
There’s more in this update too besides these highlights so check out changelog for full details!!
19/n
A number of functions have updated parameters, including `Clustered_DotPlot` which has 8!! new parameters to further customize the output plot and plot legends.
See changelog for full details on new/updated parameters in this update.
18/n
See changelog for full details on new/updated parameters in this update.
18/n
December 6, 2024 at 4:36 PM
A number of functions have updated parameters, including `Clustered_DotPlot` which has 8!! new parameters to further customize the output plot and plot legends.
See changelog for full details on new/updated parameters in this update.
18/n
See changelog for full details on new/updated parameters in this update.
18/n
scCustomize now includes `seq_zeros` function to easily create sequences with preceding zeros. You can either specify the number of preceding zeros desired or the function will set automatically based on sequence length.
17/n
17/n
December 6, 2024 at 4:36 PM
scCustomize now includes `seq_zeros` function to easily create sequences with preceding zeros. You can either specify the number of preceding zeros desired or the function will set automatically based on sequence length.
17/n
17/n
The base R `seq` family of functions has a ton of uses. However, due to the way numbers are ordered in R (and other software) it can sometimes be helpful to have preceding zeros in your number sequences in order to keep things in numerical order (e.g., 01, 02, 03, instead of 1, 2, 3).
16/n
16/n
December 6, 2024 at 4:36 PM
The base R `seq` family of functions has a ton of uses. However, due to the way numbers are ordered in R (and other software) it can sometimes be helpful to have preceding zeros in your number sequences in order to keep things in numerical order (e.g., 01, 02, 03, instead of 1, 2, 3).
16/n
16/n
Now here’s two functions that aren’t limited to scRNA-seq analysis but can generally be helpful in your work in R.
First, `Split_Vector` which allows you to split a vector in a specified number of chunks or chunks of specific length.
15/n
First, `Split_Vector` which allows you to split a vector in a specified number of chunks or chunks of specific length.
15/n
December 6, 2024 at 4:36 PM
Now here’s two functions that aren’t limited to scRNA-seq analysis but can generally be helpful in your work in R.
First, `Split_Vector` which allows you to split a vector in a specified number of chunks or chunks of specific length.
15/n
First, `Split_Vector` which allows you to split a vector in a specified number of chunks or chunks of specific length.
15/n
Now we supply that to `cells` parameter of `DoHeatmap` and the result now looks like this. We can now clearly visualize the differences in expression between the identities.
To make it even easier you can also just call `Random_Cells_Downsample` from within the `DoHeatmap` call.
14/n
To make it even easier you can also just call `Random_Cells_Downsample` from within the `DoHeatmap` call.
14/n
December 6, 2024 at 4:36 PM
Now we supply that to `cells` parameter of `DoHeatmap` and the result now looks like this. We can now clearly visualize the differences in expression between the identities.
To make it even easier you can also just call `Random_Cells_Downsample` from within the `DoHeatmap` call.
14/n
To make it even easier you can also just call `Random_Cells_Downsample` from within the `DoHeatmap` call.
14/n
However, by using `Random_Cells_Downsample` we can create a plot with equal number of cells per identity for proper visualization. Here I create a down sampled cell vector of 150 cells per group (or the max number of cells for groups smaller than 150 cells)
13/n
13/n
December 6, 2024 at 4:36 PM
However, by using `Random_Cells_Downsample` we can create a plot with equal number of cells per identity for proper visualization. Here I create a down sampled cell vector of 150 cells per group (or the max number of cells for groups smaller than 150 cells)
13/n
13/n
The default cell level heatmap scales the size of each identity based on the number of cells. This can lead to heatmaps where visualizing the expression in small clusters can be nearly impossible.
12/n
12/n
December 6, 2024 at 4:36 PM
The default cell level heatmap scales the size of each identity based on the number of cells. This can lead to heatmaps where visualizing the expression in small clusters can be nearly impossible.
12/n
12/n
This update also contains new feature called `Random_Cells_Downsample` to return randomly downsampled set of cells from object with equal numbers of cells per identity.
There are many scenarios where this can be helpful but one that I use a lot is when plotting cell-level heatmaps.
11/n
There are many scenarios where this can be helpful but one that I use a lot is when plotting cell-level heatmaps.
11/n
December 6, 2024 at 4:36 PM
This update also contains new feature called `Random_Cells_Downsample` to return randomly downsampled set of cells from object with equal numbers of cells per identity.
There are many scenarios where this can be helpful but one that I use a lot is when plotting cell-level heatmaps.
11/n
There are many scenarios where this can be helpful but one that I use a lot is when plotting cell-level heatmaps.
11/n
Just like human function, `Updated_MGI_Symbols` requires one time internet connection to download the MGI file but then stores it in local cache so not internet is required for subsequent use.
10/n
10/n
December 6, 2024 at 4:36 PM
Just like human function, `Updated_MGI_Symbols` requires one time internet connection to download the MGI file but then stores it in local cache so not internet is required for subsequent use.
10/n
10/n
The last scCustomize update brought the advance of `Updated_HGNC_Symbols` for updating human gene symbols. This update brings the equivalent function of mice using MGI database: `Updated_MGI_Symbols`.
9/n
9/n
December 6, 2024 at 4:36 PM
The last scCustomize update brought the advance of `Updated_HGNC_Symbols` for updating human gene symbols. This update brings the equivalent function of mice using MGI database: `Updated_MGI_Symbols`.
9/n
9/n
scCustomize also makes it’s first venture in spatial plotting with addition of `SpatialDimPlot_scCustom` which is sibling to `DimPlot_scCustom` and contains similar updates to color scheme and other defaults.
See new spatial vignette for details:
samuel-marsh.github.io/scCustomize/...
8/n
See new spatial vignette for details:
samuel-marsh.github.io/scCustomize/...
8/n
Plotting #5: Spatial Plotting Functions
scCustomize
samuel-marsh.github.io
December 6, 2024 at 4:36 PM
scCustomize also makes it’s first venture in spatial plotting with addition of `SpatialDimPlot_scCustom` which is sibling to `DimPlot_scCustom` and contains similar updates to color scheme and other defaults.
See new spatial vignette for details:
samuel-marsh.github.io/scCustomize/...
8/n
See new spatial vignette for details:
samuel-marsh.github.io/scCustomize/...
8/n
One new plotting function “Factor_Cor_Plot” has been added which will plot correlation between feature loadings of liger iNMF factors (works with both Seurat and Liger objects).
LIGER vignette is in progress and will I get that finished as soon as I can!
7/n
LIGER vignette is in progress and will I get that finished as soon as I can!
7/n
December 6, 2024 at 4:36 PM
One new plotting function “Factor_Cor_Plot” has been added which will plot correlation between feature loadings of liger iNMF factors (works with both Seurat and Liger objects).
LIGER vignette is in progress and will I get that finished as soon as I can!
7/n
LIGER vignette is in progress and will I get that finished as soon as I can!
7/n
Old liger functions have been updated to enable working with either old or new style liger objects.
New liger interaction functions have been added. “Subset_LIGER”, “Cells_by_Identities_LIGER”.
6/n
New liger interaction functions have been added. “Subset_LIGER”, “Cells_by_Identities_LIGER”.
6/n
December 6, 2024 at 4:36 PM
Old liger functions have been updated to enable working with either old or new style liger objects.
New liger interaction functions have been added. “Subset_LIGER”, “Cells_by_Identities_LIGER”.
6/n
New liger interaction functions have been added. “Subset_LIGER”, “Cells_by_Identities_LIGER”.
6/n
scCustomize v3.0.0 contains massive updates for interaction with new style liger objects (liger v2.0.0+). Changelog contains the full function update list but some highlights are:
Extending Seurat generic functions to work with liger (Cells, Features, Idents, WhichCells, etc)
5/n
Extending Seurat generic functions to work with liger (Cells, Features, Idents, WhichCells, etc)
5/n
December 6, 2024 at 4:36 PM
scCustomize v3.0.0 contains massive updates for interaction with new style liger objects (liger v2.0.0+). Changelog contains the full function update list but some highlights are:
Extending Seurat generic functions to work with liger (Cells, Features, Idents, WhichCells, etc)
5/n
Extending Seurat generic functions to work with liger (Cells, Features, Idents, WhichCells, etc)
5/n
In addition to new species and metrics `Add_Cell_QC` is now S3 generic function and works with either Seurat or Liger objects.
I have also added new vignette specifically dedicated to this and other QC functions in scCustomize.
samuel-marsh.github.io/scCustomize/...
4/n
I have also added new vignette specifically dedicated to this and other QC functions in scCustomize.
samuel-marsh.github.io/scCustomize/...
4/n
Object QC Functions
scCustomize
samuel-marsh.github.io
December 6, 2024 at 4:36 PM
In addition to new species and metrics `Add_Cell_QC` is now S3 generic function and works with either Seurat or Liger objects.
I have also added new vignette specifically dedicated to this and other QC functions in scCustomize.
samuel-marsh.github.io/scCustomize/...
4/n
I have also added new vignette specifically dedicated to this and other QC functions in scCustomize.
samuel-marsh.github.io/scCustomize/...
4/n
`Add_Cell_QC` also can work with ensembl IDs instead of feature names and has all ensembl IDs for default species stored within the package, no download required!
3/n
3/n
December 6, 2024 at 4:36 PM
`Add_Cell_QC` also can work with ensembl IDs instead of feature names and has all ensembl IDs for default species stored within the package, no download required!
3/n
3/n
First, QC functions.
`Add_Cell_QC_Metrics` now includes addition of % of hemoglobin counts as metric & chicken was added to default species.
Default species are now human, mouse, rat, drosophila, zebrafish, macaque, marmoset, & chicken.
One line will add all QC metrics to your object.
2/n
`Add_Cell_QC_Metrics` now includes addition of % of hemoglobin counts as metric & chicken was added to default species.
Default species are now human, mouse, rat, drosophila, zebrafish, macaque, marmoset, & chicken.
One line will add all QC metrics to your object.
2/n
December 6, 2024 at 4:36 PM
First, QC functions.
`Add_Cell_QC_Metrics` now includes addition of % of hemoglobin counts as metric & chicken was added to default species.
Default species are now human, mouse, rat, drosophila, zebrafish, macaque, marmoset, & chicken.
One line will add all QC metrics to your object.
2/n
`Add_Cell_QC_Metrics` now includes addition of % of hemoglobin counts as metric & chicken was added to default species.
Default species are now human, mouse, rat, drosophila, zebrafish, macaque, marmoset, & chicken.
One line will add all QC metrics to your object.
2/n