A revised single-cell transcriptomic atlas of Xenopus embryo reveals new differentiation dynamics

Kseniya Petrova, Maksym Tretiakov, Aleksandr Kotov, Anne H. Monsoro-Burq, Leonid Peshkin

Preprint posted on 2 January 2024 https://www.biorxiv.org/content/10.1101/2024.01.02.573882v1

A cell’s eye view of development: Kseniya Petrova et al. explore early frog development at the single-cell level, producing a new and improved atlas resource.

Selected by Rachel Mckeown

Categories: bioinformatics, cell biology, developmental biology, genetics, genomics

Background

Single-cell technologies are becoming more and more popular, providing an unprecedented level of detail which simply cannot be obtained at larger scales. The Human Cell Atlas² is the flagship project pioneering this approach in the context of human health and disease, but we cannot learn all that we need to know from studying humans alone.

To dive deeper, we turn to our trusted cohort of model organisms. When studying highly conserved early embryonic processes, the insights gained from studies using model organisms also give us a window into the development of more complex, but less accessible, organisms. Xenopus tropicalis, the Western clawed frog, is a favourite model of many developmental biologists for various reasons: the embryos are relatively large and develop rapidly, and as in all amphibians, do so externally from the mother where they can be manipulated experimentally.

To generate ideas and design experiments, researchers benefit greatly from resources such as cell atlases, which compile data of all cell types that make up an organism. These take a great deal of effort to produce but allow the worldwide research community to further probe the biggest questions of development.

In this preprint, the authors present the latest single-cell atlas for early Xenopus development and put it through its paces to demonstrate its value.

Key findings

The new atlas increases per-stage and per-cell type coverage

The raw data for sequencing was collected for the first edition of the atlas using inDrops technology⁴. This means that cells were isolated into hydrogel microspheres to undergo lysis, reverse transcription and cell-specific barcoding that labelled the mRNA. Pooled droplets then underwent reverse transcription together to generate DNA for further amplification and sequencing, with the barcodes serving as primers.

inDrops sequencing process (Version 2)⁴

Equipped with the updated Xenopus annotated genome, the researchers identified cells that were missing from .This additional information was the result of new inDrops data and better assignment to previously unannotated protein-coding genes. The authors also didn’t discard data from cells with low total gene expression, previously thought to result from mRNA loss in the experimental pipeline.

With around 50,000 more cells recovered than in the previous atlas, the number attributed to each developmental stage was increased, with each stage also benefiting from better sequencing coverage.

Figure 1: A) Number of unique molecular identifiers (UMIs) per cell at different stages B) Total number of cells per stage C) Coverage of different cell types

The authors have shared their data through the Spring web browser, with many useful practical features. Of note, rather than stamping a cell with a definite identity, they calculated, for each cell, the probability that it is every type of cell. This ‘soft’ cell type index may be more true to the biology at play in early development, where cells are still differentiating.

Comparisons with the previous atlas show improvements

The authors leveraged two complimentary machine learning approaches to evaluate how well the previous atlas annotated cell types. The kNN classified method provided ease of comparison between old and new genome annotations, while the marker-gene based approach gave more context to the cell clusters in terms of their biological roles. Overall, the previous atlas did a decent job annotating cell types, but some cell types were difficult to predict, such as those lacking a strong marker gene profile. The authors therefore re-evaluated the classification of cells in their updated atlas, getting rid of some unhelpful terms like ‘hatching gland’ and merging others, such as ‘ventral blood island’ and ‘lateral plate mesoderm’. This improved the consistency of cell types.

Using matching barcodes to transfer the annotation of cell types found in both current and former atlases, the cell type prediction accuracy was found to be fairly constant and hence robust to the updated genome annotation. However, a few new marker genes were found, and others were lost.

Updated coverage supports new cell types

Primordial germ cells (PGCs) could hardly be found in the previous atlas, with only three representative cells. They likely slipped through the net due to their low transcriptional activity during embryogenesis. However, the new atlas raised their number to 477 across multiple developmental stages, including identifiable marker genes like dnd1. The authors noted that their reads for PGCs were biased in favour of exons, suggesting that they were likely deposited by the mother as maternal transcripts are already spliced to remove their introns.

Figure 6: (A) Dotplot of the top upregulated differentially expressed genes in old (XA18) and updated (XA23) atlases – note increased marker gene presence for germ cells. B) Lower intron/exon ratio in germ cells across multiple developmental stages.

What I liked about this preprint

It amazes me at just how small a scale we can investigate development and how rapidly the technology is progressing to facilitate this. Will we ever be able to assign precise definitions and numbers to cell types? Is this even possible? Furthermore, the outcome of this paper is a user-friendly platform which is readily available to facilitate discoveries in labs across the world. As a Xenopus researcher, I fully appreciate how valuable such high-quality public data can be in designing an experiment or interpreting the results. The thriving Xenopus research community will be excited to serve as the ultimate beta testers for this tool. I look forward to seeing what new insights can be gained from having this updated cell atlas at our disposal.

Questions for the authors

Could a similar single-cell atlas be made for Xenopus laevis, the tetraploid African clawed frog? What would the challenges be?
When do you think we’ll next be due an update for the atlas? Do you think there are more cell types to discover?

References

Kseniya Petrova, Maksym Tretiakov, Aleksandr Kotov, Anne H. Monsoro-Burq, & Leonid Peshkin. A revised single-cell transcriptomic atlas of <em>Xenopus</em> embryo reveals new differentiation dynamics. bioRxiv 2024.01.02.573882 (2024) doi:10.1101/2024.01.02.573882.
Human Cell Atlas. https://www.humancellatlas.org/.
Briggs, J. A. et al. The dynamics of gene expression in vertebrate embryogenesis at single-cell resolution. Science 360, eaar5780 (2018).
Simonas Juzenas et al. inDrops-2: a flexible, versatile and cost-efficient droplet microfluidics approach for high-throughput scRNA-seq of fresh and preserved clinical samples. bioRxiv 2023.09.26.559493 (2024) doi:10.1101/2023.09.26.559493

Tags: frog

Posted on: 11 March 2024

doi: https://doi.org/10.1242/prelights.36640

Read preprint

(No Ratings Yet)

Author's response

Leonid Peshkin and Kseniia Petrova shared

Could a similar single-cell atlas be made for Xenopus laevis, the tetraploid African clawed frog? What would the challenges be?

LP: Uniquely matching the reads to a tetraploid genome is a challenge, also single-cell dissociation and barcoding might not translate to cells from species separated by 70 million years of estimated evolutionary distance.

KP: I agree with Leon, that tetraploidy will make things more difficult to analyse. Library preparation using standard droplet-based single-cell RNA-seq protocols results in reads with very biased coverage of gene body: either to 3′ UTR or to 5’UTR. So in many cases, you wouldn’t be able to distinguish between .L and .S copies of X. laevis genes. Another problem is that after duplication there must be some shifts in gene function due to reduced selection pressure. So, comparisons with other organisms that have lower ploidy will also be more difficult. But on the other side, it makes X. laevis a very exciting model for the study of evolution… and some single-cell methods allow profiling of full-length transcripts. I hope that soon there will be more such methods, more productive and less expensive.

When do you think we’ll next be due an update for the atlas? Do you think there are more cell types to discover?

LP: We are presently planning to extend the atlas to later stages, but also to profile the early stages using spatial transcriptomics. Optimistically, the new atlas should be ready in two years.

KP: We are working on many things on Xenopus and planning to do even more. (If you are interested in working with us, please send us a message). I would like to approach lineage tracing ~~if the stars ever line up in the right position in the sky~~. But as Leon says, it’s not in a short-term perspective.

Have your say Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Sign up to customise the site to your preferences and to receive alerts

Also in the bioinformatics category:

Expressive modeling and fast simulation for dynamic compartments

Till Köster, Philipp Henning, Tom Warnke, et al.

Selected by Benjamin Dominik Maier

Transcriptional profiling of human brain cortex identifies novel lncRNA-mediated networks dysregulated in amyotrophic lateral sclerosis

Alessandro Palma, Monica Ballarino

Selected by Julio Molina Pineda

Spatial transcriptomics elucidates medulla niche supporting germinal center response in myasthenia gravis thymoma

Yoshiaki Yasumizu, Makoto Kinoshita, Martin Jinye Zhang, et al.

Selected by Jessica Chevallier

Also in the cell biology category:

Fetal brain response to maternal inflammation requires microglia

Bridget Elaine LaMonica Ostrem, Nuria Dominguez Iturza, Jeffrey Stogsdill, et al.

Selected by Manuel Lessi

Alteration of long and short-term hematopoietic stem cell ratio causes myeloid-biased hematopoiesis

Katsuyuki Nishi, Taro Sakamaki, Akiomi Nagasaka, et al.

Selected by Manuel Vicente, Anaisa Ferreira

Clusters of lineage-specific genes are anchored by ZNF274 in repressive perinucleolar compartments

Martina Begnis, Julien Duc, Sandra Offner, et al.

Selected by Silvia Carvalho

Also in the developmental biology category:

Alteration of long and short-term hematopoietic stem cell ratio causes myeloid-biased hematopoiesis

Katsuyuki Nishi, Taro Sakamaki, Akiomi Nagasaka, et al.

Selected by Manuel Vicente, Anaisa Ferreira

Temporal constraints on enhancer usage shape the regulation of limb gene transcription

Raquel Rouco, Antonella Rauseo, Guillaume Sapin, et al.

Selected by María Mariner-Faulí

OGT prevents DNA demethylation and suppresses the expression of transposable elements in heterochromatin by restraining TET activity genome-wide

Hugo Sepulveda, Xiang Li, Xiaojing Yue, et al.

Selected by Mansi

Also in the genetics category:

Temporal constraints on enhancer usage shape the regulation of limb gene transcription

Raquel Rouco, Antonella Rauseo, Guillaume Sapin, et al.

Selected by María Mariner-Faulí

A long non-coding RNA at the cortex locus controls adaptive colouration in butterflies

Luca Livraghi, Joseph J. Hanly, Elizabeth Evans, et al.

AND

The ivory lncRNA regulates seasonal color patterns in buckeye butterflies

Richard A. Fandino, Noah K. Brady, Martik Chatterjee, et al.

AND

A micro-RNA drives a 100-million-year adaptive evolution of melanic patterns in butterflies and moths

Shen Tian, Tirtha Das Banerjee, Jocelyn Liang Qi Wee, et al.

Selected by Isabella Cisneros