Hacker News Reader

Comment by sixQuarks | Hacker News Reader

I'm afraid it's not the individual genes themselves that are important to life, but a specific combination of those genes. Further complicating the problem is that we have no idea how many genes comprise an "essential group". Is it a combination of 2 genes? 3, 5, 10?

When you're talking about 175 unknown genes, the combination of all of these is a huge number. It's like finding a needle in a haystack the size of the solar system.

I don't think this brute force approach is going to work, we need a different way to figure this out, but I'm confident that once figured out, it will seem simple looking back on it.

View on HN

Also, isn't there an issue of "gene" being something of an unnatural category? Nature doesn't provide a clean labeling of what part of the sequence is "a" gene, and the term historically existed (AIUI) to denote regions that the researcher found to have phenotypic relevance. So I'm not sure it's helpful to phrase a minimum in terms of "n genes". See:

https://en.wikipedia.org/wiki/Gene#Functional_definitions

This is an excellent point. It also illustrates how a useful abstraction at one level can hinder reasoning at another.

> the term historically existed (AIUI) to denote regions that the researcher found to have phenotypic relevance

That's Mendelian genetics. Geneticists focused on DNA as a physical molecule rather than a conceptual unit of inheritance use the same word to refer to a region of DNA that produces a protein.

They're both important concepts, but they're not the same concept.

Arguably, the production of a protein is nearly the smallest element of the observable characteristic of an organism produced by the interaction of its genotype with the environment, and thus sort of the fundamental unit of phenotype, so, from that perspective, they are exactly the same thing.

They are not the same thing from any useful perspective. You can remove a (protein-coding) gene, or replace it, without altering the phenotype at all. You can make changes outside protein-coding regions and get large phenotypic changes.

Okay I got the terms wrong but I think the core point stands: a gene is typically defined with respect to some research objective and area of interest, and is highly variable depending on what you're looking at. It doesn't (yet, at least) have a natural boundary. People looking at one phenomenon might see one boundary for a a gene, while others see another.

In fact, I think that term is great for comparison: "phenomenon". Asking for the "minimum genes for viable life" seems like asking for the "minimum number of natural phenomena to put a man on the moon". Are those two phenomena really just different manifestations of "gravity" or are they two things in their own right? "It depends."

I am thinking of George Boole (boolean algebra and logic) who wrote in the 18th centuary -- that a symbol (an individual gene in this case) in itself is not meaningful but that its meaning depends on how it combines with with other symbols. Quote- "..the validity of the processes of analysis does not depend upon the interpretation of the symbols which are employed, but solely on their laws of combination. Every system of interpretation which does not affect the truth of the relations supposed, is equally admissible, and it is true that the same process may,under one scheme of interpretation, represent the solution of a question on the properties of numbers, under another, that of a geometrical problem, and under a third, that of a problem of dynamics or optics." (Boole 1847)

> I don't think this brute force approach is going to work, we need a different way to figure this out, but I'm confident that once figured out, it will seem simple looking back on it.

Why not? Couldn't we use machines to automate this and analyze the results?

I think it might take a while.

http://coolconversion.com/math/factorial/What-is-the-number-...

When you're talking about 175 different genes, there are far more combinations of genes than there are atoms in the universe.

Perhaps a quantum computer can tackle this kind of problem, but nothing we have today can brute-force the solution.

They got it down to 473 genes, but let's say that minimal life is some 20 of those genes. A brute force approach would require trying 10^34 combinations[1]. And that's already with the simplifying assumption that 20 genes is the minimal set rather than 19 or 106 or some other subset.

[1] http://www.wolframalpha.com/input/?i=473+choose+20

Reply on news.ycombinator.com