Lightnews — Scholar-powered news

Lev Tsypin, PhD

@ltsyp.in

What is the world coming to?

A bag of JACK LINKS beef jerky "doritos" with "taco flavor". "Limited time offer" and $1 off!

January 11, 2025 at 2:21 AM

Lev Tsypin, PhD

@ltsyp.in

Happy new year, everyone

January 1, 2025 at 5:33 PM

Lev Tsypin, PhD

@ltsyp.in

There is clearly much more to explore in these data, as we generated over 600 co-expression clusters, with over a quarter of them being enriched for known cellular functions.

To make our analysis accessible to the community, we built standalone HTML files for interactive data exploration.

A labeled diagram of the TGNE dashboard. (A) The “Conditions Selection Tabs” allow the user to specify which life cycle phases are included within the input data to the clustering pipeline. The “Normalization Selection Tabs” allow the user to select which normalization technique should be used on the input data. (B) The search bars can be used to select genes based on their annotations. The left search bar allows searches for TTHERM_ID, common names, descriptions, and module number. The right search bar allows searches for functional annotation terms or codes. Here, “m179” was used as the search term to select the entire module that is enriched for histone-associated functional terms. (C) The heatmap representation of the normalized expression of all genes across all conditions. The selected module is highlighted, and the unselected genes are grayed out. (D) This plot shows all modules with significantly enriched functional terms, which are the same terms as those that can be searched using the right-hand search bar. Moving the cursor over any of the circles in the plot displays the enriched term, its fold-change relative to the genome background, and the Bonferroni-corrected p-value. (E) An interactive UMAP representation of the gene expression with one tab showing the UMAP embedding of each cluster and the other tab showing the UMAP embedding of each gene. Selected genes and modules are highlighted, while unselected ones are grayed out. Clicking on any circle or selecting them with one of the tools to the right of the plot selects those module(s) or gene(s) for display. (F) The graph for displaying the expression profiles of the selected genes. (G) When genes are selected, their annotation information based on the published T. thermophila genome, eggNOG, and InterProScan is populated into this table. (H) The annotation table and functional enrichment information for the selected genes/modules can be downloaded as tab-separated files using these two buttons.

December 17, 2024 at 1:06 AM

Lev Tsypin, PhD

@ltsyp.in

We find strong evidence for histone-, ribosome-, and proteasome-associated genes.

The most striking example is of the ribosome-associated genes. There is a 49-gene overlap between the microarray and RNA-seq co-expression patterns, and every one of those genes is known to be ribosomal.

Figure 4 from our preprint. The left column shows normalized co-expression patterns from the microarray dataset, and the right column shows normalized co-expression patterns from the RNA-seq dataset. The top row is for histone-associated genes, the middle row is for ribosome-associated genes, and the bottom row is for proteasome-associated genes.

December 17, 2024 at 1:06 AM

Lev Tsypin, PhD

@ltsyp.in

Every knockout had a mucocyst secretion defect! To read this figure, all you need to know is that when wildtype cells are stimulated to secrete and then are centrifuged, they end up with a mucus layer over the cell pellet. In each picture, the left tube is wildtype, and the right tube is a knockout.

Figure 3 from our preprint. In each image, the left tube shows the result of a secretion test with the wildtype strain of T. thermophila, and the right tube shows the result with a knockout strain for each of the genes listed above the photos. In each photo, the right tube has a much smaller or absent mucus layer overlying the cell pellet.

December 17, 2024 at 1:06 AM

Lev Tsypin, PhD

@ltsyp.in

Not only that, but these co-expression patterns also intersect with the list of genes that are specifically upregulated during mucocyst biogenesis!

We're publishing this mucocyst biogenesis dataset for the first time, which was collected by Prof. Lydia Bright during her PhD ~15 years ago.

Panels from figure 2 of our preprint. C and D show the normalized expression profiles of genes that are co-expressed with previously identified mucocyst biogenesis genes in both the microarray (C) and RNA-seq (D) datasets. Panel E is a volcano plot showing differential expression during mucocyst biogenesis. The red dashed lines indicate the thresholds we used for fold-change and false discovery rate, and the colors correspond to the posterior Bayesian probability of differential expression, with yellow approaching 0% and purple approaching 100%. Panel F is a Venn Diagram showing the overlap between genes that are upregulated during mucocyst biogenesis, genes that are co-expressed with mucocyst biogenesis genes in the microarray dataset, and genes that are co-expressed with mucocyst biogenesis genes in the RNA-seq dataset.

December 17, 2024 at 1:06 AM

Lev Tsypin, PhD

@ltsyp.in

After these steps, we perform simulations to test the null hypothesis that there is no way to cluster gene co-expression patterns in the data. No matter how we normalized the data, our optimized clustering was much better than random chance.

Comparison of our optimized clustering (dashed green line) against the simulated negative controls (black and purple histograms), as evidenced by a modularity metric. The left column corresponds to the microarray data, the right column corresponds to the RNA-seq data, the top row corresponds to a minimum-maximum normalization framework, and the bottom row corresponds to the z-score normalization framework. In each case, our clustering of the real data has a much higher modularity than the simulations, indicating that we are recovering true co-expression structure from the data.

Heat maps of gene expression for all genes. On the left, the heat map corresponding to the microarray dataset. On the right, the heat map corresponding to the RNA-seq dataset. Each row corresponds one of the ~19,000 genes in our dataset. The genes are sorted by cluster and inter-cluster similarity, revealing a whole-genome view of co-expression patterns. The darker the color, the closer the normalized expression is to the minimum for the given gene; the brighter the color, the closer the normalized expression is to the maximum for the given gene.

December 17, 2024 at 1:06 AM

Lev Tsypin, PhD

@ltsyp.in

T. thermophila has a complex life cycle, and the available expression datasets span bulk growth, starvation, and sexual reproduction (conjugation), as well as over a synchronized mitotic cell cycle.

But these data were collected in different labs over two decades and came from different methods.

A schematic of the T. thermophila asexual and sexual life cycles from Orias (2012). The things to notice are that the cell has two distinct nuclei, which stands in contrast to more familiar eukaryotic cells. The larger nucleus is called the "macronucleus" or the "somatic" nucleus. The smaller nucleus is called the "micronucleus" or the "germline" nucleus. The macronucleus is responsible for gene expression throughout the cell's activities, but the micronucleus only becomes relevant during sexual reproduction (conjugation). During conjugation, the micronucleus undergoes meiosis, performs reciprocal fertilization of the conjugating cells, and patterns the formation of the new macronucleus.

December 17, 2024 at 1:06 AM

Lev Tsypin, PhD

@ltsyp.in

Answering these questions requires appropriate model systems and principled approaches.

We pursued these questions in Tetrahymena thermophila, a single-celled eukaryote. Previously, we (and others) had observed that co-expressed genes in T. thermophila tend to be functionally associated.

Schematic diagram of Tetrahymena thermophila from Sachse (2014). On the left, a scanning electron micrograph of a T. thermophila cell showing its surface structure. The cell is pear-shaped and covered with hair-like cilia. On the right, a diagram of the cell, including the internal structures. Most notably, despite being a single cell, it has analogs to multicellular organisms, including an oral apparatus that it uses as a mouth.

December 17, 2024 at 1:06 AM

Lev Tsypin, PhD

@ltsyp.in

My first last-author paper is out as a preprint! 🧪

"Inferring gene-pathway associations from consolidated transcriptome datasets: an interactive gene network explorer for Tetrahymena thermophila"

I wanted to wait til bioRxiv finished formatting it, but I couldn't!

www.biorxiv.org/content/10.1...

Graphical abstract for my paper, showing a flowchart of how we consolidated microarray and RNA-seq gene expression data to generate informative co-expression clusters. The steps are: alignment to the newest genome model, gene filtering and quality control, expression normalization and averaging, clustering the high-dimensional dataset, performing parameter scans and computational negative controls, and building an interactive tool for data exploration.

December 17, 2024 at 1:06 AM

Lev Tsypin, PhD

@ltsyp.in

"My other car is a..."

Photo of a car with the license plate "HOENEST"

December 15, 2024 at 11:45 PM

Lev Tsypin, PhD

@ltsyp.in

My latest pots :)

The same two bowls from the top view. The left bowl's bottom has blue overglaze, and the right bowl's bottom has red overglaze.

December 2, 2024 at 5:35 AM

Lev Tsypin, PhD

@ltsyp.in

Finally, Microsoft. I'm disappointed with the update. Before, they had a dying protist (note the blebbing) poisoned by bacteria it had eaten (the little rods). These might be Chromobacterium violaceum, which makes a bluish toxin.

Now, it's just a worse version of what Google and Twitter have.

The 2020 Microsoft emoji. The blue protist (note the circle nucleus and pseudopodia) is on a dark, bubbly background, reminiscent of the blebbing process of eukaryotic cells decaying. Inside the protist are three rods that are eroding its membrane. The rods are darker blue, like Chromobacterium violaceum. The charisma of this scene is staggering.

Welp, the new Microsoft emoji is just a goofy green cell with some sparse protrusions and a smattering of blue circles in the middle. There is no clear membrane. It's not clear whether the blue circles are meant to be internal or external to the cell: they might just be stains on the surface, for all we know.

November 16, 2024 at 4:19 AM

Lev Tsypin, PhD

@ltsyp.in

WhatsApp stuck with their (great!) bacterium. It looks Gram-positive (note the dark purple hue of the stain). The cell appendages are too short to be pili or flagellae, but their number and placement is good. There appears to be DNA, but no compartments, so it's easy to determine it bacterial.

WhatsApp's microbe is bright purple, which is reminiscent of the Gram stain, which usually distinguishes between bacteria with one and two membranes based on the thickness of their peptidoglycan cell wall. Here, we see the rod-shaped cell with many short appendages sticking out like tentacles all around it. In the middle of the cell are some circles that I am choosing to interpret as DNA.

November 16, 2024 at 4:19 AM

Lev Tsypin, PhD

@ltsyp.in

Twitter's microbe is... nothing particularly charismatic. It's unchanged from 2020, it's probably another amoeboid eukaryote, and it doesn't hold a candle to the Facebook one.

I'm all about protists! But I am also all about microbial diversity 😤

Twitter's microbe is almost identical to Google's new one. It has a single membrane, a green color scheme, and some vesicles with a potential nucleus in a darker shade. It also has pseudopodia, so I'd also call it amoeboid.

November 16, 2024 at 4:19 AM

Lev Tsypin, PhD

@ltsyp.in

Apple's emoji remains the same, and still salient. Look at that, it's a coronavirus! Is a virion a microbe? ¯\_(ツ)_/¯

The Apple "microbe" emoji is an illustration of a coronavirus. It's green, and it's not very detailed, but the spike proteins are pretty good.

November 16, 2024 at 4:19 AM

Lev Tsypin, PhD

@ltsyp.in

Now, Facebook. The design is unchanged, and for good reason. This appears to be an amoeboid cell, with evidence of photosynthetic organelles. If it had a shell (and fewer plastids) I'd guess Paulinella, but this more likely represents a Chlorarachniophyte. A very deep cut 👨‍🍳👌

The Facebook emoji for a microbe is a plump, squishy little guy. In some ways, it's similar to the new Google emoji, but the details that are emphasized are much more charismatic. It has a single membrane, and it also appears to be ciliated. Without a scale bar, it's hard to be sure, but it smacks of eukaryote to me. But crucially, all throughout the cell, there are green organelles, which immediately look like plastids to me.

A microscopic image of a real Chlorarachniophyte for comparison. The photosynethic organelles are probably taken from eaten green algae in a process called kleptoplasty!

Image taken from https://alchetron.com/Chlorarachniophyte#chlorarachniophyte-f826a535-517f-4741-89c3-ac1df0f4418-resize-750.jpg.

November 16, 2024 at 4:19 AM

Lev Tsypin, PhD

@ltsyp.in

First up, Google. Last time, I appreciated the diversity and details, like the polyphosphate granules in the lower left cell, but I disqualified them for including a centipede.

Someone realized the error, and their update is much more reasonable: an amoeboid eukaryote! Look at those pseudopodia! 😍

Google's 2020 emoji. There are five "microbes," each of a different color. Two appear to have nuclei, which would imply that they are eukaryotes, but they are of the same size as the ones that are intended to be bacteria. While there are very small eukaryotes, this is ridiculous. In the top left corner is a critter that looks like a centipede more than anything else. In the bottom left corner is a compelling little guy with granules that are phase-separated from the rest of the cytoplasm, and which do not cross the inner membrane. If they had just this one, it would be excellent.

This is a single microbe, a variant of the middle image in the 2020 emoji. There is a distinct nucleus, what appear to be vesicles or phagosomes, and pretty reasonable cilia all around the cell. Good job!

November 16, 2024 at 4:19 AM

Lev Tsypin, PhD

@ltsyp.in

Perhaps you've heard of comparative genomics, but have you heard of comparative emojinomics? 🧪

In 2020, I rated the most common microbe emoji on their microbiality and charisma. Two of the designs have been updated since then, so it's time to revisit this important task.

A grid showing the "microbe" emoji from Facebook, Google, Microsoft, Apple, Twitter, and WhatsApp. Each is discussed individually in the thread below.

November 16, 2024 at 4:19 AM

Lev Tsypin, PhD

@ltsyp.in

Another fun thing to do, though it requires patience, is to make a Winogradsky column. Here's an example of how you can do it: uwaterloo.ca/science/wino.... The most important thing is never open it indoors because it'll be the worst thing you've ever smelled. I know from experience 😅