Similarity of human and chimp DNA is down.

[Zaius137]

Blayz...

You should get a clue when someone is letting you off the hook.

“It's simple math. If I have 62.7% of 30BP matches then i have 30throot of 0.627 (=0.9841) of 1BP matches, and then 0.9841^20 for 20BP matches and 0.9841^50 for 50BP matches.”

Pray tell what are you demonstrating? Probability? Binomial expansion over an interval? What is this? The .9841 seems to be occurrences/interval but you don’t deal with it in that way. Is it the total mutations/total bp? If it is why are you raising it to the interval power? By the way the selected interval is taken into 10,000 evenly spaced random entry points in a comparison of equivalent segments (those that align); the author just happened to use 30bp which is a multiple of 120bp commonly used in genetics. You have yet to justify your calculation…

As I stated early on, I am not here to win an argument. Remember I mentioned that the people who are here for that purpose don’t ever learn much. I never learn anything from someone who just makes stuff up.

 

[Zaius137]

Blayz....

Is this you on Youtube?

http://www.youtube.com/watch?v=Fl6eDoCRrkw&noredirect=1

 

[sfs]

“It's simple math. If I have 62.7% of 30BP matches then i have 30throot of 0.627 (=0.9841) of 1BP matches, and then 0.9841^20 for 20BP matches and 0.9841^50 for 50BP matches.”

Pray tell what are you demonstrating? Probability? Binomial expansion over an interval? What is this? The .9841 seems to be occurrences/interval but you don’t deal with it in that way. Is it the total mutations/total bp?

No, it's the probability of not having a mutation in a single base.

If it is why are you raising it to the interval power?

Because that's how you calculate the probability of not having a mutation in the interval. This is completely trivial math.

By the way the selected interval is taken into 10,000 evenly spaced random entry points in a comparison of equivalent segments (those that align);

A fact that matters not at all for this discussion.

the author just happened to use 30bp which is a multiple of 120bp commonly used in genetics.

The last time I looked at this thread, you were claiming the author had a reason for using 30 bp. Have you given up on that now? So why pick 30 bp, since you'll get a different answer for 20 bp or for 40 bp? (And what is that 120 bp that's supposed to be commonly used in genetics? Where on earth did that come from?)

As I stated early on, I am not here to win an argument.

Perhaps you should start trying.

Remember I mentioned that the people who are here for that purpose don’t ever learn much. I never learn anything from someone who just makes stuff up.

Have you learned yet how to calculate the probability that a 30 bp segment of sequence will have 0 mutations if 1.4% of bases have mutations?

 

[Loudmouth]

Zaius137,

If I compare a 30 bp segment of chimp and human DNA and they differ by one base substitution what is the percent similarity between the two sequences? This is a very simple question, can you answer it?

 

[Loudmouth]

it is 62.7% real matches (I will not continually do the math again and again).

What constitutes a real match? What sequence database is the author using?

our genome only compares segment for segment 62.7% of the time;

How do you choose the length of the segment? Why not 10 or 100 bp?

I let you go on without examining one of the major problems I brought up in the beginning; the 2.4 Gb/ 3.1Gb align able segments which you do not address.

It has been addressed ad nauseum. Contigs were thrown out if it was impossible to determine where the pieces fit in the overall sequence. Only 2.4 Gb was deemed quality sequence. There was alignment between chimp and human DNA. The problem was that the alignment was not contiguous or it was ambiguous due to poor overlaps.

The absolute best claim the evolutions view could possibly make would be a 2.4/3.1 ~ 77% similarity.

You have yet to tell us how why the other 23% of the genome would have no homology to human DNA. How did you determine this?

Also the additions of base pairs or subtraction of base pairs is not ethical in trying to have the genes come out to the same length.

No such thing was done.

I pointed out the fact several times that there can be 6 separate coding of a single gene by altering the entry point in both directions; by adding one bp the coding losses its context (it is not the same gene).

And? How do you determine if there was an indel until you do the actual sequencing? You only look for ORF's after you have the sequencing done, not before.

You better explain your reasoning here because I see no foundation for this. At minimum I need a citation (see if you can quantify it without making stuff up).

When using 1 bp segments the authors of the chimp genome paper got about 98% similarity. If you used a 3.1 Gb segment you would get 0% similarity, even between humans (excluding identical twins). It is quite obvious that 10 bp segments would return a similarity between 1 bp and 30 bp segments, and anything above 30 bp segments would return a lower percentage of similarity. By adjusting the length of the segment you can return whatever percent similarity you want, which makes it arbitrary.

In effect, you are arguing that the distance between the floor and the basketball hoop has increased since it is 120 inches, not the 10 feet that most people report.

 

[Zaius137]

If the percentage similarity holds for a single bp this simple math would work. But we are discussing a global similarity of 98.4% of genes. Just very the percentage by 1% which is within the uncertainty of the match.

(.99)^50 = .61
(.98)^50 = .36

One says 61% the other says 36%. Your conclusion using this method is worthless.

The only way to do a comparison is with a wild match of bp similarity. You must use the poisson distribution to generalize the interval. You made this mistake earlier.

The 30bp is a good statistical offset as verified in the secular and non-secular findings.

 

[Loudmouth]

But we are discussing a global similarity of 98.4% of genes.

How are those genes compared? Base to base or 30 bp segment to 30 bp segment?

Just very the percentage by 1% which is within the uncertainty of the match.

(.99)^50 = .61
(.98)^50 = .36

One says 61% the other says 36%. Your conclusion using this method is worthless.

What sfs was calculating was the probability of two orthologous 30 bp segments having 100% identity given the ~98% similarity on a base to base comparison. As it turns out, the two methods produce comparable similarities.

What you keep forgetting is the units. You are arguing that a 30 bp comparison should return the same results as a 1 bp comparison. This is obviously wrong.

The only way to do a comparison is with a wild match of bp similarity.

Why? Why can't you compare orthologous sequence and look for the number of bases that differ along that string? That is the accepted method in all the genome papers I have read. I have not read one where they compare 30 bp segments.

You must use the poisson distribution to generalize the interval.

Interval of what?

The 30bp is a good statistical offset as verified in the secular and non-secular findings.

The offset was random for each segment chosen. The author used a random number generator to indicate the start of the 30 bp segment. He then used a BLAST search to find an exact match for that 30 bp segment in the other genome. If BLAST did not find an exact match he counted it as a mismatch, even if there was sequence that was one base pair different. That is how he did it. If the genomes are ~98% different using a 1 bp comparison then you should observe that ~65% of 30 bp segments should have 100% identity between the two genomes.

Again, you are arguing that 120 inches is different than 10 feet simply because the numbers are different.

 

[Zaius137]

“It is better to keep your mouth shut and appear stupid than to open it and remove all doubt”
Mark Twain

Was a very wise man… I could vary with his opinion on religion though.

Yes Loudmouth they do add and remove bps to aid in matching…

“In addition to mismatches, the chimpanzee and human genomes also differ in their lengths. When aligning any two sequences, it is occasionally necessary to insert or omit a nucleotide (or more) in order to maintain the best alignment.”

http://documents.clubexpress.com/doc...yArVOo%2FgM%3D

I have already made my point about the unaligned segments. You still can not tell me why you can ignore them scientifically.

I have read all the quoted articles in this thread and can not vary my opinion in the smallest degree.

 

[Loudmouth]

Yes Loudmouth they do add and remove bps to aid in matching…

“In addition to mismatches, the chimpanzee and human genomes also differ in their lengths. When aligning any two sequences, it is occasionally necessary to insert or omit a nucleotide (or more) in order to maintain the best alignment.”

http://documents.clubexpress.com/doc...yArVOo%2FgM%3D

They are talking about the ALIGNMENT, not the sequence. Do you understand the difference or not?

sequence 1: AAAAATAAAAA
sequence 2: AAAAA-AAAAA

In the comparison above a gap was added to sequence 2 in order to produce an alignment. This does not change the reported sequence for 1 or 2. Instead, the T is reported as an indel, a site were either a T was inserted in one species (seq 1) or deleted in the other species (seq 2). Without analysis of other species within the framework of a phylogeny it is difficult to tell which has occurred, an insertion or deletion.

Just so we are clear, the reported sequence for sequence 2 is not AAAAA-AAAAA, it is the same sequence without the dash.

This is even more apparent if you had quoted the next three sentences from that non-reviewed paper:

"Evolutionary biologists assume that such adjustments are necessary because nucleotides have been inserted or deleted over the course of evolution. Since I do not accept the common ancestry of humans and chimpanzees, I will refer to these adjustments as "gaps," following the tradition of computational biology (e.g. Altschul et al. 1997; Pearson 1998). The chimpanzee/human genome alignment contains approximately ten million gaps, covering 67 million nucleotides."

It is easy to see that he is referring to GAPS that are put in in order to compare the sequences. He is NOT talking about random bases being thrown in to fill those gaps. Therefore, the author agrees that indels within coding regions resulting in frame shift mutations would have been detected by the chimp genome consortium.

I have already made my point about the unaligned segments. You still can not tell me why you can ignore them scientifically.

You don't even understand how alignments are done, so your point is based on an ignorance of the process. You can have a contig that is very similar to human DNA, but that contig will be thrown out if it has poor overlap and/or a small section of the contig also aligns with a different portion of the genome. These contigs were thrown out because of ambiguous reconstruction, not because of a lack of similarity to human DNA.

I have read all the quoted articles in this thread and can not vary my opinion in the smallest degree.

Your dogma is noted.

 

[Zaius137]

Loudmouth...

If you add a bp or remove a bp to a gene that codes in multiple directions then you change the context for possible 6 proteins. It is intellectually dishonest.



I need to put some thing to rest. If you compare bp to bp what do you get? Not 90% or anything in-between you either get 100% match or no match not 98.74%. A 1bp method showing 90% is ludicrous. The main reason why a bp to bp does not work. Thanks (sfs) I just realized this fact when working some numbers.

 

[Blayz]

As I stated early on, I am not here to win an argument.

You are succeeding admirably at this goal.

Blayz....

Is this you on Youtube?

You got me. I only pretend to be a 46 yr old Australian living in Singapore. In reality I am a 30 something American and like superbowling, apparently. That's when they turn on the disco lights over the lanes and hand out cash prizes for strikes.

I need to put some thing to rest. If you compare bp to bp what do you get? Not 90% or anything in-between you either get 100% match or no match not 98.74%

Riiight, so if I have 10 bp and 9 out of 10 match then that is either 100% or nothing, not 90%? I am beginning to see where your problems lie.

If you add a bp or remove a bp to a gene that codes in multiple directions then you change the context for possible 6 proteins. It is intellectually dishonest.

coding segments make up about 2% of the genome. indels in the remaining 98% do not cause frameshifts because there are no frames to shift.

the author just happened to use 30bp which is a multiple of 120bp commonly used in genetics. You have yet to justify your calculation…

As I stated early on, I am not here to win an argument. Remember I mentioned that the people who are here for that purpose don’t ever learn much. I never learn anything from someone who just makes stuff up.

shaudenfreudengasmic

 

[Nostromo]

If you add a bp or remove a bp to a gene that codes in multiple directions then you change the context for possible 6 proteins. It is intellectually dishonest.

You're saying that if you have two identical segments e.g. 200 bases long and you insert one new base at the beginning, that whole segment is now a 0% match because it is functionally different rather a different sequence?

 

[sfs]

If the percentage similarity holds for a single bp this simple math would work. But we are discussing a global similarity of 98.4% of genes.

Sorry, but we're not discussing anything about the similarity of genes. Every comparison you've raised -- the ones in the genome paper, the crappy Wiki article, and the pointless UD article -- has been a sequence comparisons, not a comparison of gene.

Just very the percentage by 1% which is within the uncertainty of the match.

It's not clear what, if anything, this statement means.

(.99)^50 = .61
(.98)^50 = .36

Yes, different per-base differences will give different 50-bp sequence match rates. What's your point?

One says 61% the other says 36%. Your conclusion using this method is worthless.

Let me try to understand here. Because you can make up two numbers out of the air, and they're different, that means that the actual numbers we measured are worthless? Or are you saying that the UD method would have gotten a different number if the per-base difference were something else? What are you saying?

The only way to do a comparison is with a wild match of bp similarity.

No, that's a really dumb way to do the comparison. Note that neither you nor the UD guy who carried out this test has yet offered a single cogent reason for using this approach. Occasionally you'll offer some vague explanation, but then back off it when you're challenged.

You must use the poisson distribution to generalize the interval. You made this mistake earlier.

Why would you use a Poisson distribution for something that is binomially distributed?

The 30bp is a good statistical offset as verified in the secular and non-secular findings.

The 30 bp method is a pretty dumb approach with zero statistical justification; simply repeating that it's statistically sound isn't going to make it smell any better. It has been "verified" by nothing at all. All that's been verified is that it gives the result you'd expect if the genomes were actually as diverged as we said they were. Why anyone would use this approach is still a deep mystery.

 

[sfs]

I need to put some thing to rest. If you compare bp to bp what do you get? Not 90% or anything in-between you either get 100% match or no match not 98.74%. A 1bp method showing 90% is ludicrous. The main reason why a bp to bp does not work. Thanks (sfs) I just realized this fact when working some numbers.

If you compare bp to bp, you do indeed get either 0% or 100% match at each site. Ignoring indels for the moment, what you get is a 100% match 98.7% of the time and a 0% match 1.3% of the time. Pretty straightforward method, I'd say.

 

[Naraoia]

A good idea. You could even download the human and chimp genomes from NCBI, and the perl script from the uncommon descent author. Said script contains a variable currently set to 30 which could be changed to 20 or 50.

EDIT: I'd do a simulation with a higher sample rate that generates 300-400 million sequences, then map with bowtie, but I am guessing Zaius doesn't have access to a decent 64bit linux machine.

Hehe, I was actually trying to code something smaller-scale in R, seeing as R is the only programming language I still have a functional knowledge of. Then I found that R and I... had different ideas of what you can do with vectors, and just scrapped the whole thing to save me some headache. I think my programming skills need an upgrade

I totally missed there was a script on UD...

 

[Loudmouth]

Loudmouth...

If you add a bp or remove a bp to a gene that codes in multiple directions then you change the context for possible 6 proteins.

This has nothing to do with the actual sequence. No one is adding any BASES. Thye are adding gaps in the alignment process to detect indels. The actual reported sequence does not have these gaps. The gaps are added by the algorithm for the purposes of comparison ONLY, not for determining the sequence. It is intellectually dishonest to claim otherwise.

You can not do an alignment until you have the sequence. You can not detect indels until you have the sequence from both genomes. No one is adding BASES to the reported chimp sequence. To claim otherwise is intellectually dishonest.

If you compare bp to bp what do you get? Not 90% or anything in-between you either get 100% match or no match not 98.74%.

The 98.74% is the AVERAGE across the genome. For example:

seq 1: AAAAAAAAAA
seq 2: AAAAAAAAAT

The AVERAGE similarity across those two 10 bp sequences is 90% because 9 out of 10 bases match. If I were doing a 10 bp comparison instead of a 1 bp comparison the similarity would be 0%. Do you understand this or not? Why is this so hard to understand?

How do you think you get a 65% similarity with 30 bp comparisons? It is either 0% or 100%, right? How can it be 65% similarity? Perhaps you should take up this argument with the author over at UD.

 

[Zaius137]

I am up against the evolutionist brick wall.

“The 98.74% is the AVERAGE across the genome. For example:

seq 1: AAAAAAAAAA
seq 2: AAAAAAAAAT

The AVERAGE similarity across those two 10 bp sequences is 90% because 9 out of 10 bases match. If I were doing a 10 bp comparison instead of a 1 bp comparison the similarity would be 0%. Do you understand this or not? Why is this so hard to understand?”

How would you get 98.74% over 100bp. Where did the .74 come from? It is because a 1bp interval does not work. Yes you are comparing bp to bp in your example above. How does an “A” compare to an “A”, how does an “A” compare to a “T”, 98%? Why don’t you understand your analysis of the statistical comparison is bogus…

 

[cupid dave]

Did it ever occur to you fellows that the 22 links our Paleontologists say are the 22 species in our ascent to modern man are the same idea as the 22 names in Genesis from Adam through the three racial stocks of Ham, Japeth, and Shem?


Gen 5:25-32

 

[Blayz]

I am up against the evolutionist brick wall.

Yes, it divides creation land from reality.

How would you get 98.74% over 100bp. Where did the .74

He didn't. In his example he used 90%.

What if I had 9874 exact matches over 10 000bp. Whip out your calculator and work out the average then.

Yes you are comparing bp to bp in your example above. How does an “A” compare to an “A”

We call that a match

how does an “A” compare to a “T”

We call that a mismatch.

statistical

Please stop using words you do not understand. I mean, you clearly do not understand the concept of "average", and that is basic math, a long way from statistics.

 

[sfs]

Did it ever occur to you fellows that the 22 links our Paleontologists say are the 22 species in our ascent to modern man are the same idea as the 22 names in Genesis from Adam through the three racial stocks of Ham, Japeth, and Shem?


Gen 5:25-32

Well, no, I've never thought of that. Perhaps because there aren't 22 particular species that stand out among our ancestors, and perhaps because as an explanation it doesn't really make much sense. Are you suggesting that the Adam of Genesis went about on four legs and had a brain the size of a chimpanzee?