Combining datasets after pre-processing

Author(s) AvatarWendi Bacon AvatarJonathan Manning
Editor(s) orcid logoAvatarHelena Rasche
Tester(s) AvatarJulia Jakiela
Overview
Questions:

  • I have some AnnData files from different samples that I want to combine into a single file. How can I combine these and label them within the object?

Objectives:

  • Combine data matrices from different samples in the same experiment

  • Label the metadata for downstream processing

Requirements:
Time estimation: 1 hour
Supporting Materials:
Last modification: Sep 28, 2022
License: Tutorial Content is licensed under Creative Commons Attribution 4.0 International License The GTN Framework is licensed under MIT

Introduction

This tutorial will take you from the multiple AnnData outputs of the previous tutorial to a single, combined AnnData object, ready for all the fun downstream processing. We will also look at how to add in metadata (for instance, SEX or GENOTYPE) for analysis later on.

Agenda

In this tutorial, we will cover:

  1. Introduction
    1. Get Data
    2. Concatenating objects
  2. Adding batch metadata
  3. Mitochondrial reads
  4. Pulling single cell data from public resources
  5. Conclusion

Get Data

The sample data is a subset of the reads in a mouse dataset of fetal growth restriction Bacon et al. 2018 (see the study in Single Cell Expression Atlas here and the project submission here). Each of the 7 samples (N701 –> N707) has been run through the workflow from the Alevin tutorial.

You can access the data for this tutorial in multiple ways:

  1. Your own history - If you’re feeling confident that you successfully ran a workflow on all 7 samples from the previous tutorial, and that your resulting 7 AnnData objects look right (you can compare with the answer key history), then you can use those! To avoid a million-line history, I recommend dragging the resultant datasets into a fresh history

    There 3 ways to copy datasets between histories

    1. From the original history

      1. Click on the icon (History options) on the top of the history panel
      2. Click on Copy Dataset

      3. Select the desired files

      4. Give a relevant name to the “New history”

      5. Click on the new history name in the green box that have just appear to switch to this history

    2. From the View all histories

      1. Click on View all histories on the top right
      2. Switch to the history in which the dataset should be copied
      3. Drag the dataset to copy from its original history
      4. Drop it in the target history

    3. From the target history

      1. Click on User in the top bar
      2. Click on Datasets
      3. Search for the dataset to copy
      4. Click on it
      5. Click on Copy to History

  2. Importing from a history - You can import this history

    1. Open the link to the shared history
    2. Click on the Import history button on the top right
    3. Enter a title for the new history
    4. Click on Import

  3. Uploading from Zenodo (see below)

Hands-on: Data upload for 7 files
  1. Create a new history for this tutorial (if you’re not importing the history above)

  2. Import the different AnnData files and the experimental design table from Zenodo.

    https://zenodo.org/record/7075635/files/Experimental_Design.tsv
https://zenodo.org/record/7075635/files/N701-400k.h5ad
https://zenodo.org/record/7075635/files/N702-400k.h5ad
https://zenodo.org/record/7075635/files/N703-400k.h5ad
https://zenodo.org/record/7075635/files/N704-400k.h5ad
https://zenodo.org/record/7075635/files/N705-400k.h5ad
https://zenodo.org/record/7075635/files/N706-400k.h5ad
https://zenodo.org/record/7075635/files/N707-400k.h5ad
    • Copy the link location

    • Open the Galaxy Upload Manager ( on the top-right of the tool panel)

    • Select Paste/Fetch Data

    • Paste the link into the text field

    • Press Start

    • Close the window
  3. Rename the datasets

  4. Check that the datatype is h5ad, otherwise you will need to change each file to h5ad!

    • Click on the pencil icon for the dataset to edit its attributes
    • In the central panel, click on the Datatypes tab on the top
    • Select datatypes
      • tip: you can start typing the datatype into the field to filter the dropdown menu
    • Click the Save button

Inspect the Experimental Design text file. This shows you how each N70X corresponds to a sample, and whether that sample was from a male or female. This will be important metadata to add to our sample, which we will add very similarly to how you added the gene_name and mito metadata previously!

Concatenating objects

Hands-on: Concatenating AnnData objects
  1. Manipulate AnnData with the following parameters:
    • “Annotated data matrix”: N701-400k
    • “Function to manipulate the object”: Concatenate along the observations axis
    • “Annotated data matrix to add”: Select all the other matrix files from bottom to top, N707 to N702
    Warning: Warning!: N707 to N702!

    You are adding files to N701, so do not add N701 to itself!

    • “Join method”: Intersection of variables
    • “Key to add the batch annotation to obs”: batch
    • “Separator to join the existing index names with the batch category”: -
  2. Rename output Combined Object

Now let’s look at what we’ve done! Unfortunately, AnnData objects are quite complicated, so the won’t help us too much here. Instead, we’re going to use a tool to look into our object from now on.

Hands-on: Inspecting AnnData Objects
  1. Inspect AnnData with the following parameters:
    • “Annotated data matrix”: Combined object
    • “What to inspect?”: General information about the object
  2. Inspect AnnData with the following parameters:
    • “Annotated data matrix”: Combined object
    • “What to inspect?”: Key-indexed observations annotation (obs)
  3. Inspect AnnData with the following parameters:
    • “Annotated data matrix”: Combined object
    • “What to inspect?”: Key-indexed annotation of variables/features (var)

Now have a look at the three Inspect AnnData outputs.

Question
  1. How many cells do you have now?
  2. Where is batch information stored?
  1. If you look at the General information output, you can see there are now 338 cells, as the matrix is now 338 cells x 35734 genes. You can see this as well in the obs (cells) and var (genes) file sizes.
  2. Under Key-indexed observations annotation (obs). Different versions of the Manipulate tool will put the batch columns in different locations. The tool version in this course puts batch in the 8th column. Batch refers to the order in which the matrices were added. The files are added from the bottom of the history upwards, so be careful how you set up your histories when running this (i.e. if your first dataset is N703 and the second is N701, the batch will call N703 0 and N701 1!)

Adding batch metadata

I set up the example history with the earliest indices at the bottom.

The files are numbered such that dataset #1 is N701, dataset #2 is N702, etc., up through #7 as N707. This puts N707 at the top and N701 at the bottom in the Galaxy history.
Figure 1: Correct history ordering for combining datasets in order

Therefore, when it is all concatenated together, the batch appears as follows:

Index Batch Genotype Sex
N701 0 wildtype male
N702 1 knockout male
N703 2 knockout female
N704 3 wildtype male
N705 4 wildtype male
N706 5 wildtype male
N707 6 knockout male

If you used Zenodo to import files, they may not have imported in order (i.e. N701 to N707, ascending). In that case, you will need to tweak the parameters of the next tools appropriately to label your batches correctly!

The two critical pieces of metadata in this experiment are sex and genotype. I will later want to color my cell plots by these parameters, so I want to add them in now!

Hands-on: Labelling sex
  1. Replace Text in a specific column with the following parameters:
    • “File to process”: output of Inspect AnnData: Key-indexed observations annotation (obs) )

    • “1. Replacement”

      • “in column”: Column: 8 - or whichever column batch is in
      • “Find pattern”: 0|1|3|4|5|6
      • “Replace with”: male
    • + Insert Replacement

    • “2. Replacement”

      • “in column”: Column: 8
      • “Find pattern”: 2
      • “Replace with”: female
    • + Insert Replacement

    • “3. Replacement”

      • “in column”: Column: 8
      • “Find pattern”: batch
      • “Replace with”: sex

    Now we want only the column containing the sex information - we will ultimately add this into the cell annotation in the AnnData object.

  2. Cut columns from a table with the following parameters:
    • “Cut columns”: c8
    • “Delimited by”: Tab
    • “From”: output of Replace text
  3. Rename output Sex metadata

That was so fun, let’s do it all again but for genotype!

Hands-on: Labelling genotype
  1. Replace Text in a specific column with the following parameters:
    • “File to process”: output of Inspect AnnData: Key-indexed observations annotation (obs)

    • “1. Replacement”

      • “in column”: Column: 8
      • “Find pattern”: 0|3|4|5
      • “Replace with”: wildtype
    • + Insert Replacement

    • “2. Replacement”

      • “in column”: Column: 8
      • “Find pattern”: 1|2|6
      • “Replace with”: knockout
    • + Insert Replacement

    • “3. Replacement”

      • “in column”: Column: 8
      • “Find pattern”: batch
      • “Replace with”: genotype

    Now we want only the column containing the genotype information - we will ultimately add this into the cell annotation in the AnnData object.

  2. Cut columns from a table with the following parameters:
    • “Cut columns”: c8
    • “Delimited by”: Tab
    • “From”: output of Replace text
  3. Rename output Genotype metadata

You might want to do this with all sorts of different metadata - which labs handled the samples, which days they were run, etc. Once you’ve added all your metadata columns, we can add them together before plugging them into the AnnData object itself.

Hands-on: Combining metadata columns
  1. Paste two files side by side with the following parameters:
    • “Paste”: Genotype metadata
    • “and”: Sex metadata
    • “Delimit by”: Tab
  2. Rename output Cell Metadata

Let’s add it to the AnnData object!

Hands-on: Adding metadata to AnnData object
  1. Manipulate AnnData with the following parameters:
    • “Annotated data matrix”: Combined object
    • “Function to manipulate the object”: Add new annotation(s) for observations or variables
    • “What to annotate?”: Observations (obs)`
    • “Table with new annotations”: Cell Metadata

Woohoo! We’re there! You can run an Inspect AnnData to check now, but I want to clean up this AnnData object just a bit more first. It would be a lot nicer if ‘batch’ meant something, rather than ‘the order in which the Manipulate AnnData tool added my datasets’.

Hands-on: Labelling batches
  1. Manipulate AnnData with the following parameters:
    • “Annotated data matrix”: output of Manipulate AnnData - Add new annotations
    • “Function to manipulate the object”: Rename categories of annotation
    • “Key for observations or variables annotation”: batch
    • “Comma-separated list of new categories”: N701,N702,N703,N704,N705,N706,N707
  2. Rename output Batched Object

Huzzah! We are JUST about there. However, while we’ve been focussing on our cell metadata (sample, batch, genotype, etc.) to relabel the ‘observations’ in our object…

Mitochondrial reads

Do you remember when we mentioned mitochondria early on in this tutorial? And how often in single cell samples, mitochondrial RNA is often an indicator of stress during dissociation? We should probably do something with our column of true/false in the gene annotation that tells us information about the cells. You will need to do this whether you have combined FASTQ files or are analysing just one.

Hands-on: Calculating mitochondrial RNA in cells
  1. AnnData Operations with the following parameters:
    • “Input object in hdf5 AnnData format”: Batched Object
    • “Format of output object”: AnnData format
    • “Gene symbols field in AnnData”: NA.
    • “Flag genes that start with these names”: Insert Flag genes that start with these names
    • “Starts with”: True
    • “Var name”: mito
  2. Rename output Annotated Object

Well done! I strongly suggest have a play with the Inspect AnnData on your final Pre-processed object to see the wealth of information that has been added. You are now ready to move along to further filtering! There is a cheat that may save you time in the future though…

Pulling single cell data from public resources

If you happen to be interested in analysing publicly available data, particularly from the Single Cell Expression Atlas, you may be interested in the following tool Moreno et al. 2020 which combines all these steps into one! For this tutorial, the dataset can be seen here with experiment id of E-MTAB-6945.

Hands-on: Retrieving data from Single Cell Expression Atlas
  1. EBI SCXA Data Retrieval with the following parameters:
    • “SC-Atlas experiment accession”: E-MTAB-6945
    • “Choose the type of matrix to download”: Raw filtered counts

    Now we need to transform this into an AnnData objects

  2. Scanpy Read10x with the following parameters:
    • “Expression matrix in sparse matrix format (.mtx)”: EBI SCXA Data Retrieval on E-MTAB-6945 matrix.mtx (Raw filtered counts)
    • “Gene table”: EBI SCXA Data Retrieval on E-MTAB-6945 genes.tsv (Raw filtered counts)
    • “Barcode/cell table”: EBI SCXA Data Retrieval on E-MTAB-6945 barcodes.tsv (Raw filtered counts)
    • “Cell metadata table”: EBI SCXA Data Retrieval on E-MTAB-6945 exp_design.tsv

It’s important to note that this matrix is processed somewhat through the SCXA pipeline, which is quite similar to this tutorial, and it contains any and all metadata provided by their pipeline as well as the authors (for instance, more cell or gene annotations).

Conclusion

Combining data files.
Figure 2: Workflow - Combining datasets

You’ve reached the end of this session! You may be interested in seeing an example history and workflow. Note that the workflow will require changing of the column containing the batch metadata depending on how you are running it. The final object containing the total the reads can be found in here.

To discuss with like-minded scientists, join our Gitter channel for all things Galaxy-single cell! Gitter.

Key points

  • Create a single scanpy-accessible AnnData object from multiple AnnData files, including relevant cell metadata according to the study design

  • Retreive partially analysed data from a public repository

  • Ready for the next tutorial in this case study

Frequently Asked Questions

Have questions about this tutorial? Check out the tutorial FAQ page or the FAQ page for the Transcriptomics topic to see if your question is listed there. If not, please ask your question on the GTN Gitter Channel or the Galaxy Help Forum

Useful literature

Further information, including links to documentation and original publications, regarding the tools, analysis techniques and the interpretation of results described in this tutorial can be found here.

References

  1. Bacon, W. A., R. S. Hamilton, Z. Yu, J. Kieckbusch, D. Hawkes et al., 2018 Single-Cell Analysis Identifies Thymic Maturation Delay in Growth-Restricted Neonatal Mice. Frontiers in Immunology 9: 10.3389/fimmu.2018.02523
  2. Moreno, P., N. Huang, J. R. Manning, S. Mohammed, A. Solovyev et al., 2020 User-friendly, scalable tools and workflows for single-cell analysis. bioRxiv. 10.1101/2020.04.08.032698 https://www.biorxiv.org/content/early/2020/04/09/2020.04.08.032698

Feedback

Did you use this material as an instructor? Feel free to give us feedback on how it went.
Did you use this material as a learner or student? Click the form below to leave feedback.

Click here to load Google feedback frame

Citing this Tutorial

  1. , 2022 Combining datasets after pre-processing (Galaxy Training Materials). https://training.galaxyproject.org/training-material/topics/transcriptomics/tutorials/scrna-case_alevin-combine-datasets/tutorial.html Online; accessed TODAY
  2. Batut et al., 2018 Community-Driven Data Analysis Training for Biology Cell Systems 10.1016/j.cels.2018.05.012 


@misc{transcriptomics-scrna-case_alevin-combine-datasets,
author = "Wendi Bacon and Jonathan Manning",
title = "Combining datasets after pre-processing (Galaxy Training Materials)",
year = "2022",
month = "09",
day = "28"
url = "\url{https://training.galaxyproject.org/training-material/topics/transcriptomics/tutorials/scrna-case_alevin-combine-datasets/tutorial.html}",
note = "[Online; accessed TODAY]"
}
@article{Batut_2018,
    doi = {10.1016/j.cels.2018.05.012},
    url = {https://doi.org/10.1016%2Fj.cels.2018.05.012},
    year = 2018,
    month = {jun},
    publisher = {Elsevier {BV}},
    volume = {6},
    number = {6},
    pages = {752--758.e1},
    author = {B{\'{e}}r{\'{e}}nice Batut and Saskia Hiltemann and Andrea Bagnacani and Dannon Baker and Vivek Bhardwaj and Clemens Blank and Anthony Bretaudeau and Loraine Brillet-Gu{\'{e}}guen and Martin {\v{C}}ech and John Chilton and Dave Clements and Olivia Doppelt-Azeroual and Anika Erxleben and Mallory Ann Freeberg and Simon Gladman and Youri Hoogstrate and Hans-Rudolf Hotz and Torsten Houwaart and Pratik Jagtap and Delphine Larivi{\`{e}}re and Gildas Le Corguill{\'{e}} and Thomas Manke and Fabien Mareuil and Fidel Ram{\'{\i}}rez and Devon Ryan and Florian Christoph Sigloch and Nicola Soranzo and Joachim Wolff and Pavankumar Videm and Markus Wolfien and Aisanjiang Wubuli and Dilmurat Yusuf and James Taylor and Rolf Backofen and Anton Nekrutenko and Björn Grüning},
    title = {Community-Driven Data Analysis Training for Biology},
    journal = {Cell Systems}
}

Congratulations on successfully completing this tutorial!