Nikolai Slavov
banner
slavov-n.bsky.social
Nikolai Slavov
@slavov-n.bsky.social
Mentor, scientist & engineer. Having fun in @slavovlab.bsky.social and Parallel Squared Technology Institute @parallelsq.bsky.social with biology & single-cell proteomics.

https://nikolai.slavovlab.net
Pinned
A basic intro with some strategies how to avoid mistakes when using dimensionality reduction.

2/n

youtu.be/nIB-WjWzDSc?...
Introduction to dimensionality reduction | Statistics for proteomics
YouTube video by Nikolai Slavov
youtu.be
Among all proteome changes in Alzheimer’s disease (AD), 20S proteasomes stand out.

⬛️ Their abundance strongly declines in AD patients.

=> This decline parallels a reduction in 20S substrate proteins and leads to abnormal protein accumulation in AD.

www.biorxiv.org/content/10.1...
November 19, 2025 at 1:38 PM
Protein regulation varies across single cells, cell types, tissues, and organs.

The brain proteome is highly sculpted by protein degradation.

Why ?

This is our model: biorxiv.org/content/10.1...
November 17, 2025 at 1:14 PM
How do plants orient towards light ?

In the absence of obvious physical sensing organs like lenses, how do plants work out the precise direction from which light is coming ?
November 16, 2025 at 12:34 PM
Fitness influencers promote super-high-protein diets, and protein-fortified foods and protein supplements form a market worth tens of billions of USD.

Yet, studies show there’s only so much protein the body can use.
November 15, 2025 at 4:17 PM
Dimensionality reduction can see structures that do not exist and miss structures that exist.

𝐓𝐡𝐞 𝐬𝐢𝐦𝐩𝐥𝐞𝐬𝐭 𝐞𝐱𝐩𝐥𝐚𝐧𝐚𝐭𝐢𝐨𝐧 𝐢𝐬𝐧’𝐭 𝐚𝐥𝐰𝐚𝐲𝐬 𝐭𝐡𝐞 𝐛𝐞𝐬𝐭 𝐨𝐧𝐞.

1/n
November 14, 2025 at 12:45 PM
Current efforts focus on predicting transcriptional responses to gene silencing.

Even if spectacularly successful, they leave open key questions:
◾️Functions of genetic mutations & polymorphisms
◾️Proteoform functions
◾️Protein abundance regulation

Biology is complex.
November 13, 2025 at 3:39 PM
The protein and mRNA abundance voronoi diagrams offer remarkably different perspectives.

Each tile in the maps represents a gene product; its size is proportional to the fraction of the gene product in hepatocytes.
November 12, 2025 at 12:53 PM
A reminder that the abundance of 'housekeeping' proteins varies significantly across cells and tissues.

GAPDH is often slandered as a 'housekeeping' protein, and its abundance varies significantly across human tissues.

How do you define a 'housekeeping' protein ?
November 11, 2025 at 11:56 AM
The protein concentration in the cytoplasm is so high that the average protein has a water hydration shell with a thickness of only ≈ 10 water molecules separating it from the adjacent protein hydration shell.

1/2
November 10, 2025 at 12:57 PM
"Some participants have already found a way to hack the score, exposing fundamental flaws in how bio-foundation models are evaluated."

𝐄𝐯𝐚𝐥𝐮𝐚𝐭𝐢𝐧𝐠 𝐦𝐨𝐝𝐞𝐥𝐬 𝐢𝐬 𝐡𝐚𝐫𝐝. 𝐈𝐭 𝐧𝐞𝐞𝐝𝐬 𝐦𝐨𝐫𝐞 𝐚𝐭𝐭𝐞𝐧𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐫𝐞𝐬𝐨𝐮𝐫𝐜𝐞𝐬.

1/2
November 9, 2025 at 4:08 PM
Reposted by Nikolai Slavov
Protein degradation plays major roles in aging physiology & pathology.

We can start analyzing it quantitatively at higher resolution & scale.

A promising new direction for aging research 🚀
Aging is associated with changes in protein abundance in many organs and tissues.

We find that the protein changes are cell-type specific, and thus muted in previous bulk analysis.

Protein clearance rates change with age and drive the ...

www.biorxiv.org/content/10.1...
October 30, 2025 at 2:52 PM
A beautiful example illustrating a difference between plasma protein analysis by different technologies.

Peptides from some proteoforms correlate better to epitope-based measurements (Olink) than others.

The measurements differ because they reflect different proteoforms !
November 8, 2025 at 2:03 PM
Reposted by Nikolai Slavov
The majority of isoform pairs (of alternatively spliced transcripts) share less than 50% of their interactions.

In the context of interactome networks, alternative isoforms tend to behave like distinct proteins rather than minor variants of each other.
Proteoforms encoded by the same gene have different interactomes.

This challenges:
1⃣ The interpretation of data from affinity reagents directed towards shared epitopes.

2⃣ The assignment of functions to genes.

When will biological research focus on such functional differences?
November 6, 2025 at 1:41 PM
Proteoforms encoded by the same gene have different interactomes.

This challenges:
1⃣ The interpretation of data from affinity reagents directed towards shared epitopes.

2⃣ The assignment of functions to genes.

When will biological research focus on such functional differences?
November 6, 2025 at 12:52 PM
People are talking about the “language of cells” and the “language of biology”, merging science and poetry.

I continue to prefer rigorously defined terminology.

Metaphors can be beautiful, appropriate, and useful, but rigorous science is better served by nonfiction prose.
November 5, 2025 at 12:47 PM
My commute to PTI.
November 3, 2025 at 2:36 PM
As an undergrad at MIT, I attended many VC presentations. They hammered home the same point:

✅ The most important success factor is the team.

In the ensuing decades, I've seen evidence for it over and over.

I much prefer a great team over a pile of cash.
November 3, 2025 at 2:20 PM
Discuss.

What are the implications for student learning and evaluation.
November 2, 2025 at 6:19 PM
George Boole was born exactly 210 years ago today. He was largely self‑taught, the son of a shoemaker, yet founded what we now call Boolean algebra.
November 2, 2025 at 6:04 PM
𝐶𝑜𝑚𝑏𝑖𝑛𝑎𝑡𝑜𝑟𝑖𝑎𝑙 scaling may unlock a new level of throughput. This is enabling a long-awaited experiment !

Our team, led by Jason Derks, enabled combinatorial-scaling of mass spec proteomics throughput by developing multiplexing in the time domain, timePlex.

1/n
November 1, 2025 at 4:59 PM
One gene can code for proteoforms with:
◼️ different functions
◼️ different sequences
◼️ different regulation
◼️ different cell type expression

Such examples should inform the interpretation of gene deletion experiments seeking to infer "gene functions".
November 1, 2025 at 1:51 PM
What are effective ways to increase the rigor of scientific research and results reporting ?
November 1, 2025 at 12:09 PM
Nascent proteins may be phosphorylated during translation & the phosphosite buried in the protein’s inner core

Such buried phosphosite are found in ~ 1/3 of the phosphorylated human proteins.

They can influence protein abundance (👇), providing another example for proteoform specific regulation.
October 31, 2025 at 10:31 AM
Reposted by Nikolai Slavov
@slavov-n.bsky.social and @anshulkundaje.bsky.social on current trends of escalating hype in #AntiScience
October 30, 2025 at 1:55 PM