Oolon Colluphid

Hi Vinnie, good to have you back, especially in a new improved (apparently) non-creationist edition!

Error of understanding No1: sure, the chances of a mutation occurring at a particular point in a particular genome are miniscule. But the thing about probabilities is that it depends on opportunities, on the number of tries.

The chances of getting a 6 in a single dice throw is 1:6. I’m not about to try and calculate it, but the chances of not getting a six in, say, a million throws must be close to 0.

So there’s two additional factors the writer has overlooked: the size of the genome (number of places there could be a mutation), and especially relevant here, the size of the population -- the number of copies of gene X there are around that could get miscopied. Remember, too, that this second factor is also spread through time: it’s not just the number of copies in the present generation, but also down through the generations. This is why bacteria can evolve so rapidly to antibiotics. They have large populations, and a very rapid population turnover. A million bacteria can rack up a thousand generations in (by our standards) next to no time.

Genomes are generally rather large; most populations contain many individuals; and down time, there are lots of generations. These three things cut the odds of a mutation in general to roughly 1, and vastly improve the chances of a copying error turning up where ‘required’. I gather that each human contains three or four mutations, differences from their parents’ DNA.

Another point on numbers: in sexual organisms, gametes -- the germ-line cells -- are produced in large quantities: another opportunity for copying errors.

If mutations are so rare, how come Down Syndrome and achondroplastic dwarfism are common enough for most people to have seen and even know several examples? Downs is an entire extra haploid copy of a chromosome. Over 80% of people with achondroplasia (caused by a mutation in the FGFR3 gene, chromosomal locus 4p16) have it as a result of a new gene mutation.

1. ‘Rarer’, qualified as above....

2. ‘Beneficial’ depends entirely on the environment (including the rest of the genome) it occurs in. A gene for sharper teeth might be beneficial in an environment including meat eating, meat digestion, running after prey etc, but be of less use in a herbivore. Most mutations are in fact neutral, for a variety of reasons (see any good genetics textbook).

A specific example: there is a human genetic disorder, phenylketonuria, caused by a mutation in the gene that produces phenylalanine hydroxylase, which breaks down excesses (beyond bodily needs) of this amino acid. As homozygous individuals develop through infancy, the excess phenylalanine leads to mental retardation. If such an individual is brought up in an environment free of phenylalanine, however, the trait does not show itself. Thus if the normal human diet were different, it would not be disadvantageous.

Since the environment determines what's beneficial, neutral or harmful, if the environment changes, a previously neutral mutation may become beneficial (or harmful). Suppose our carnivore has blunter or more crushing back teeth. This may be neutral if it eats mainly meat; if it (ie the lineage) starts eating more plant matter, these teeth may become advantageous, and so something selection can pick up on. Think pandas. Similarly, achondroplasia does not otherwise affect fertility: if the environment is suitable, living to adulthood and reproducing is no problem. This is especially clear in the many dog species which are achondroplastic: corgis, dacshunds, pomeranians etc. The inclusion in their environment of human aesthetics has turned a presumably disadvantageous mutation into a beneficial one -- lineages with it prosper.

This does not follow at all. Evolution builds on what there already is; everything alive today already has ancestors every one of which ‘made it’ through life’s hurdles. Not all of these ancestors’ contemporaries did. So what we’ve got is stuff that’s already very good at being alive and getting to reproduce. So there’s no obvious reason why a mutation would not pass down generations. It may take a few for even a disadvantageous one to be ‘spotted’ by natural selection.

What mutations do is simply add to the variation in the gene pool of the population. It is populations and lineages that evolve, remember: evolution is a change in the frequency of genes in a population across time. If the environment changes, what’s needed to survive in it may well change, and different traits can come to the fore.

Eventually the crux: the Argument from Personal Incredulity and the Tornado in a Junkyard Strawman.

How many mutations? Millions, of course. But spread out over a vast amount of time.

The odds? Nowhere near as huge as protrayed, because the odds are the odds of each step, not of all of it at once. The odds do not stack up into implausibility, because at each round of the game you only keep the ones that work. Dawkins has illustrated it something like this. Get a theatre audience to all stand up and flip a coin. Get those that get tails to sit. That’s about half of them. Repeat. And repeat. In an audience of 500, it’ll take about seven or eight goes to have just one or two people left, those who got a string of heads. The odds of someone flipping eight heads in a row are 1 in 256. Not good odds... but by only keeping in the game the ones who did get heads at each try, a few such ‘lucky’ folks will be found.

The odds of each step, of course, are merely 1 in 2.

Could an eye evolve in one step? Of course not... but nobody except creationists think that’s what evolution suggests happened. Could it have evolved from something very slightly different -- say, a slightly less clear lens, a rod cell or two fewer? The answer must surely be yes (and such variation must surely be apparent in the population). Could that eye have arisen by a single step from something slightly different again? Logically, it has to be yes, provided the step is small enough to not be overly improbable. And so on, back and back, provided each step is small enough to cut its odds.

Hope that helps.

Cheers, Oolon

[ May 24, 2002: Message edited by: Oolon Colluphid ]</p>

RufusAtticus

Well for 100,000 independent throws the probability of not getting at least one 6 is

7.50576974701612106096973318606778854070363111850
7662486640070473798753732283513835037267067004014
9603274178839030161412278848964916655017447947025
9976398237856193157715617297947670240215822000527
0445091070172280107277551459451568167867292146140
4401046685055502583186357733491230766631782634229
6516435627580557762134984733254831849446168668046
4865253500662766557979193126844065982168997211897
3408661436256676702048137387847037989424670604491
0794109841124820565735149754067866759578382402957
79544941196777232698962 x 10^-7919

Multiply that by itself 10 times and you the probablilty in a million trials.

~~RvFvS~~

[edited by scigirl to try and fix formatting]
[ May 24, 2002: Message edited by: scigirl ]
{Me too - Pantera}
[ May 24, 2002: Message edited by: Pantera ]
{Edited by RvFvS - because people touched my post. }

[ May 25, 2002: Message edited by: RufusAtticus ]</p>

Oolon Colluphid

Thanks Rufus. (How d'you do that? )

Since that looks confusing on the screen, it means roughly 0.000000000000000000000000000000000000000075, but with going on eight thousand zeros between the point and the seven, instead of just those forty...

So, uh, the chances of not getting a six are about 0 then?

Oolon

Coragyps

Or, in layman's terms, about the same as the chances of Kent Hovind getting a clue.

nogods4me

Actually IMHO, it is Hovind's followers who are in need of a clue. It seems obvious to me that Hovind knows he is a liar. He is not a poor misguided, ignorant person. He is a lying, deceitful, fraud of a man.

Coragyps

ng4m - you're right - I just was going for the cheap shot.

RufusAtticus

Oolon,

For any one event the probability of not geting a six is (5/6). Since each event is independent the probability in X events is, (5/6)^X.

To calculate it, I put (5/6)^100000 in MS's powercalc for WinXP with 512 bits of precision, and out came that answer.

~~RvFvS~~

seebs

The best way to understand it is this:

Deal a hand of bridge. Pick up one of the hands of cards. Ask whether it looks like it could have formed "randomly".

Now, calculate the probability of getting those cards in that order by dealing 13 cards from a deck. It'll be 1/54*1/53*..., and it will be *VERY* small.

Wow! Must be a miracle, huh?

Mageth

Yeah, especially since you're dealing from a 54-card deck. My decks all have 52 cards. Must be a mutation.

seebs

Oh, give me a break. You know Christians can't count.