Apes and humans have different designs

mindlight · Mar 15, 2013

sfs said:
You don't have to assume anything about phylogeny or evolution. The question is simply, how similar are two pieces of DNA? We can decide what similarity implies later (or not at all). According to Tomkins's approach, the two pieces in Loudmouth's post are only 60% similar. Do you really think that's a sensible measure of similarity?

Or consider two pages from two different books. Suppose they each have 1000 characters, and are identical except that one has an extra space in the middle of the page. According to Tomkins, those two pages are 50% similar. In fact, if that page is in the middle of the book, and every other page is identical, then the two books are still 50% similar, if you happen to be looking at the whole book. Seriously, do you think this is a good measure of similarity?

I think I find the computer code analogy better to the book analogy. Compare 2 pages of text on which a couple of words are different on one page rather than the other or a few are missing and I grant you the difference is .3% or whatever. But with the computer code analogy referencing different objects in a single line of code will produce an entirely different result to the point where I could say that even a single letters difference between 2 pages of otherwise identical code made the 2 sets of code incompatible and in that case I would be tempted to say there was a 0% alignment. Also one cannot just consider the single pages. Add up these differences over a whole book and you have an entirely different story. For example one story refers to Jim and another to sally. The story takes on a completely different meaning with just a single word changed with the implications of gender etc then thrown in.

sfs · Mar 15, 2013

mindlight said:
So you have assumed that the gaps in the alignment are not significant.

I've done nothing of the sort. I haven't assumed anything about the importance of any of the differences. Some of the gaps probably are key to why humans and chimpanzees are different, and some probably do nothing at all. As I just wrote in another post, all I'm doing is assessing how well Tomkins is measuring DNA similarity, and I think he's doing a terrible job.

Here's the background: when my colleagues sequenced the chimpanzee genome, they did a very detailed and careful comparison with the human genome. They found that there were 35 million places (1.23%) where the two genomes differed by a single base (a single substitution), and another 5 million places where they differed by an insertion or deletion; the 5 million indels added up to about 1.5% unique sequence in each genome. From those findings you could come up with various numbers to describe how different the human genome is from the chimpanzee genome, but 2.7% seems like a reasonable value to use. Using almost exactly the same data, Tomkins is saying that the two genomes are really 30% different (and, by the way, my colleagues are bunch of conniving low-lifes). But the measure he's using is simply ridiculous.

Also (as a matter of interest, and I am asking cause I do not know the answer)- what is the significance of the unknown sequences(N's)? Are these not yet understood, unobservable or what?

I don't sequence genomes, but I can think of at least three classes of problem that could be involved.
1) Large blocks of DNA that is highly repetitive, i.e. short blocks of the same DNA repeated thousands of times. It's very difficult to figure out exactly how much of it there is, or where the slight variations lie in the sea of sameness. Much of this is structurally important -- the centromeres and telomeres -- but the specific sequence doesn't carry much information, so missing those regions in the sequence probably isn't too serious. (There can be genes tucked away in some of them, and they have determined their sequence -- they just can't figure out where they go in the assembly.)
2) Other repetitive DNA. This is the result of various duplications, including transposable elements that insert copies of themselves into the genome. In some cases it may not be easy to distinguish one copy from another, quite similar one.
3) Low coverage. Some spots in the genome simply weren't sequenced enough times for the sequencers to get an accurate read on the content there. This would not be an issue for the human genome reference sequence, which was done to very high depth, but would be for parts of the chimpanzee genome.

sfs · Mar 15, 2013

mindlight said:
I think I find the computer code analogy better to the book analogy. Compare 2 pages of text on which a couple of words are different on one page rather than the other or a few are missing and I grant you the difference is .3% or whatever. But with the computer code analogy referencing different objects in a single line of code will produce an entirely different result to the point where I could say that even a single letters difference between 2 pages of otherwise identical code made the 2 sets of code incompatible and in that case I would be tempted to say there was a 0% alignment. Also one cannot just consider the single pages. Add up these differences over a whole book and you have an entirely different story. For example one story refers to Jim and another to sally. The story takes on a completely different meaning with just a single word changed with the implications of gender etc then thrown in.

As I've noted, that is a possible argument, but it's not the one Tomkins is making. He is strictly and only talking about the degree of similarity between DNA sequence in this study.

(As an aside, I'll note that I spend my professional life writing computer code and analyzing DNA, and I'd say that the analogy between the two is not a very good one for understanding how DNA works, and especially not for how it responds to mutation.)

ThouShaltNotPoe · Mar 15, 2013

mindlight said:
I think I find the computer code analogy better to the book analogy. Compare 2 pages of text on which a couple of words are different on one page rather than the other or a few are missing and I grant you the difference is .3% or whatever. But with the computer code analogy referencing different objects in a single line of code will produce an entirely different result to the point where I could say that even a single letters difference between 2 pages of otherwise identical code made the 2 sets of code incompatible and in that case I would be tempted to say there was a 0% alignment. Also one cannot just consider the single pages. Add up these differences over a whole book and you have an entirely different story. For example one story refers to Jim and another to sally. The story takes on a completely different meaning with just a single word changed with the implications of gender etc then thrown in.

I've talked with many young earth creationists (of whom I was once a member of that group for many years) and it seems that many like to talk about PERCENTAGES of similarities and differences between genomes. BUT WHY DO THEY THINK PERCENTAGES ARE THE MOST IMPORTANT THING? I'm far more interested in the NESTED HIERARCHIES OF THE CODE.

Percentages make a very poor and not-so-meaningful measure of COMMON DESCENT VS. COMMON DESIGN. But once you start looking and the nested hierarchies in the data, then COMMON DESCENT looks NOTHING like the COMMON DESIGN! (Yet, most creationists pretend that the two look alike.)

But here is another misleading percentage that I've often noticed groups like Creation.com lie about. They post articles with statements like, "What the evolutionists don't tell you is that some of their most important fossils of ancient humans/hominids represent only 38% of the total skeleton." But anybody who know anatomy knows that 38% is a LOT OF WONDERFUL INFORMATION. But the average creationist critic has no idea that BILATERAL SYMMETRY can mean that a 50% skeleton remaining can be virtually a COMPLETE ANATOMICAL SPECIMEN. After all, the right side of the human body is identical to the left side in terms of BONES. And so a 38% remnant of a skeleton ---depending upon WHICH portions add up to the 38% --- can mean that scientists have as much as 3/4 of what they need to study the skeletal anatomy!

And, of course, there are other kinds of symmetry among living organisms wherein having only a small percentage of the total skeleton may be an impressive find indeed.

I just find this a good example where PERCENTAGES just aren't very meaningful unless you know the requirements of the field of study involved!

sfs · Mar 15, 2013

Update:
Well, ignore my previous results on ungapped alignments. While looking for an example of a gapped alignment to show mindlight this morning, I noticed something really odd: if I run BLAST on each query individually, rather than as a single batch job as I had been doing, all of the missing matches reappear. Rather than 68% success at matching, it's now 100%. (Note that this effect only seems to occur with ungapped alignments.) I corresponded with someone at NCBI who handles questions about BLAST, and he or she confirmed the behavior and that it appears to be a bug in the program.

So how does the comparison look now? If I use as my measure of sequence similarity the quantity (number of identical aligned bases) / (length of query sequence), I find that the mean similarity between human and chimp (for 300 base reads) is 89.7% for ungapped alignements, while the similarity for gapped alignments is 97.8%. That Tomkins got a significantly lower value makes me suspect that he was bitten by the same BLAST bug, but I couldn't say for certain.

For my own amusement, and because I've been kind of laid up sick today and too stupid to do much productive, I plotted the position of the my ~500 randomly chosen sequences in the two genomes, as determined by BLAST. That is, one axis is where the sequence was in chimpanzee chromosome 12 and the other is where BLAST says the corresponding sequence is in human chromosome 12. Here's what they look like (this is for the gapped alignments):

ungapped_sm
You can see the handful of scattered incorrect matches (which are also the ones with crappy scores and large number of mismatches). You can also see the massive pericentric inversion that makes the center of the chromosome run in different directions in the two species.

AV1611VET · Mar 15, 2013

ThouShaltNotPoe said:
BUT WHY DO THEY THINK PERCENTAGES ARE THE MOST IMPORTANT THING? I'm far more interested in the NESTED HIERARCHIES OF THE CODE.

I already explained that with ontological reductionism.

Or do you just like to yell?

Instead of looking at us like we're all simians, try looking at us like we're snap-on parts.

Or do you just like to yell?

mindlight · Mar 16, 2013

ThouShaltNotPoe said:
I've talked with many young earth creationists (of whom I was once a member of that group for many years) and it seems that many like to talk about PERCENTAGES of similarities and differences between genomes. BUT WHY DO THEY THINK PERCENTAGES ARE THE MOST IMPORTANT THING? I'm far more interested in the NESTED HIERARCHIES OF THE CODE.

Percentages make a very poor and not-so-meaningful measure of COMMON DESCENT VS. COMMON DESIGN. But once you start looking and the nested hierarchies in the data, then COMMON DESCENT looks NOTHING like the COMMON DESIGN! (Yet, most creationists pretend that the two look alike.)

But here is another misleading percentage that I've often noticed groups like Creation.com lie about. They post articles with statements like, "What the evolutionists don't tell you is that some of their most important fossils of ancient humans/hominids represent only 38% of the total skeleton." But anybody who know anatomy knows that 38% is a LOT OF WONDERFUL INFORMATION. But the average creationist critic has no idea that BILATERAL SYMMETRY can mean that a 50% skeleton remaining can be virtually a COMPLETE ANATOMICAL SPECIMEN. After all, the right side of the human body is identical to the left side in terms of BONES. And so a 38% remnant of a skeleton ---depending upon WHICH portions add up to the 38% --- can mean that scientists have as much as 3/4 of what they need to study the skeletal anatomy!

And, of course, there are other kinds of symmetry among living organisms wherein having only a small percentage of the total skeleton may be an impressive find indeed.

I just find this a good example where PERCENTAGES just aren't very meaningful unless you know the requirements of the field of study involved!

Actually the % is less important than the significance and effect of the difference. 99.7% is not a 100% and can produce an entirely different result. We are using numbers to express the fact of difference. If a single difference in a sequence of DNA changes everything then it is 100% NOT optimally aligned. So your point is a MOOT one.

Also as AV1611VET suggests we are not just our genes- and the ontological reduction in that assertion is the real methodological problem here.

Eight Foot Manchild · Mar 16, 2013

mindlight said:
My focus is on something way beyond human so anthropomorphizing is the wrong word.

You are ascribing the human qualities of personality and intent to a natural phenomenon. That is the exact definition of anthropomorphizing.

The fact that you imagine these qualities to also be possessed by your god does not magically lend credibility to your assertion. It's still the same fallacy.

mindlight · Mar 16, 2013

sfs said:
As I've noted, that is a possible argument, but it's not the one Tomkins is making. He is strictly and only talking about the degree of similarity between DNA sequence in this study.

(As an aside, I'll note that I spend my professional life writing computer code and analyzing DNA, and I'd say that the analogy between the two is not a very good one for understanding how DNA works, and especially not for how it responds to mutation.)

That's an honest answer within the limits you have set and it is interesting to read what you post.

mindlight · Mar 16, 2013

sfs said:
Update:
Well, ignore my previous results on ungapped alignments. While looking for an example of a gapped alignment to show mindlight this morning, I noticed something really odd: if I run BLAST on each query individually, rather than as a single batch job as I had been doing, all of the missing matches reappear. Rather than 68% success at matching, it's now 100%. (Note that this effect only seems to occur with ungapped alignments.) I corresponded with someone at NCBI who handles questions about BLAST, and he or she confirmed the behavior and that it appears to be a bug in the program.

So how does the comparison look now? If I use as my measure of sequence similarity the quantity (number of identical aligned bases) / (length of query sequence), I find that the mean similarity between human and chimp (for 300 base reads) is 89.7% for ungapped alignements, while the similarity for gapped alignments is 97.8%. That Tomkins got a significantly lower value makes me suspect that he was bitten by the same BLAST bug, but I couldn't say for certain.

For my own amusement, and because I've been kind of laid up sick today and too stupid to do much productive, I plotted the position of the my ~500 randomly chosen sequences in the two genomes, as determined by BLAST. That is, one axis is where the sequence was in chimpanzee chromosome 12 and the other is where BLAST says the corresponding sequence is in human chromosome 12. Here's what they look like (this is for the gapped alignments):

ungapped_sm
You can see the handful of scattered incorrect matches (which are also the ones with crappy scores and large number of mismatches). You can also see the massive pericentric inversion that makes the center of the chromosome run in different directions in the two species.

Very interesting - thanks for that. What language is BLAST written in and is it open source and available or is it a proprietary programme only for researchers?

I was interested to read your comments and corrections to how individual queries sequenced within a batch job ("the quantity (number of identical aligned bases) / (length of query sequence)")- That seems to be a crucial question - how is the query on the human DNA related to the correct query on the chimpanzee DNA? By doing an individual query you probably bypass any indexing system used to reference the 2 to one another but in a sense it is already obvious the 2 will match given that you have chosen 2 sequences of the same length or whatever to match to each other. How would it deal with a query in the sequence which was substantially different in actual composition but was also of the same length etc.

89.7% - 97.8% is not as dramatic as what Tomkins said but there is still a difference. If the difference makes all the difference then we can still see a different design for chimps and humans, which is I suppose obvious since both are significantly different in practice also even not allowing for the ontological reductionism of the naturalistic methodologies employed here.

mindlight · Mar 16, 2013

Found the BLAST programme

Download BLAST Software and Databases Documentation

Tomkins used:
ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/

BLAST: Basic Local Alignment Search Tool

BLAST - Wikipedia, the free encyclopedia

ftp://hgdownload.cse.ucsc.edu/goldenPath/panTro2/chromosomes/

http://hgdownload.cse.ucsc.edu/goldenPath/hg19/chromosomes/

BLASTN algorithm parameters for the main study were as follows: -word_size 11, -evalue 10, -max_target_seqs 1, -dust no, -soft_masking false, -ungapped.

mindlight · Mar 16, 2013

mindlight said:
Found the BLAST programme

Download BLAST Software and Databases Documentation

Tomkins used:
ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/

BLAST: Basic Local Alignment Search Tool

BLAST - Wikipedia, the free encyclopedia

ftp://hgdownload.cse.ucsc.edu/goldenPath/panTro2/chromosomes/

Index of /goldenPath/hg19/chromosomes

BLASTN algorithm parameters for the main study were as follows: -word_size 11, -evalue 10, -max_target_seqs 1, -dust no, -soft_masking false, -ungapped.

I wanted to assess the validity of this experiment and its results and as a learning aid to what is a very interesting field of study to me. As a computer programmer myself I have been amazed at how accessible the scientific methodologies and procedures used in this experiment were to me. Previously I had expected this to be the preserve of experts and something to leave to them in fact. Also from a personal point of view I am interested in how the area of computing is interacting with DNA analysis.

So some comments given some extra reading (see above for reading list)

The size of the "words" used to obtain matches seems significant and given that BLASTN gives a score to mapped matches one has also to ask if these matches are occurring in the right places e.g A query from chromosone 12 should correspond with target results from chromosone 12. So the chromosone by chromosone analysis done by Tomkins seems quite sensible. Excluding gaps also seems wise as the conventional wisdom is that in the evaluation of the HSP score by Gumbel extreme value distribution local alignments containing gaps are not proven. Leaving in the gaps also seems to lead to too many false positives from low complexity areas e.g. Ns- Nucleic acid sequences and X's -protein sequences. Also it appears that alignments can occur even where some letters in a sequence are actually different. The programme gives a score to a high quality match but this may not be a precise one
GLKFAGKLAFF may match with GLKFFGKLAFF for instance even though these are not the same. The heuristic scoring approach used by BLASTN makes a significant speed difference compared to FASTA or SMITH-WATERMAN or some other bespoke non open source solutions but these generally regarded to be at the expense of accuracy so it might have been better to use another programme. BLASTN is only measuring alignments not the chimp DNA that has got nothing to do with human DNA and vice versa (unanchored sequence contigs). Some of this it seems is regarded as junk DNA by many scientists but as Tomkins points out this is definitely not the case. Also large gaps in DNA sequence alignment are being ommitted so the reported levels of similarity end up being larger as a result.

So in short to prove a 99.7% alignment between the kind of stuff BLASTN will identify for the analysis and to do so with the methods BLASTN uses is to miss a large chunk of stuff that is not compared and is to say things are statistically aligned when this is not a precise match and significant features of the sequencing arrangement are ommitted from the analysis.

Tomkins tried to deal with many of these flaws with the program and assumptions made by it and other programmes with his analysis and the way he constructed his assumptions. Larger words (11 letters were used) , X & N gaps were excluded, only one match was allowed for each alignment, chromosones were done one by one rather than seeking matches across the whole db and sequence matching was optimised so that the right sequences were compared. It was noted that large amounts of DNA could not be matched at all cause they were not in the overlap between Chimps and Humans.

So overall his conclusion seems quite sound:

Tomkins said:
Only 69% of the chimpanzee X chromosome was similar to human and only 43% of the Y chromosome. Chimp autosomal similarity to human on average was 70.7% with a range of 66.1% to 77.9%, depending on the chromosome (Table 1 and Fig. 1). Genome-wide, only 70% of the chimpanzee DNA was similar to human under the most optimal sequence-slice conditions.
Chimpanzees and humans share many localized protein-coding regions of high similarity. However, overall there is extreme DNA sequence discontinuity between the two genomes. The current study along with several other recent reports confirm this. This defies standard evolutionary time-scales and dogmatic presuppositions about a common ancestor.

Loudmouth · Mar 19, 2013

mindlight said:
Excluding gaps also seems wise as the conventional wisdom is that in the evaluation of the HSP score by Gumbel extreme value distribution local alignments containing gaps are not proven. Leaving in the gaps also seems to lead to too many false positives from low complexity areas e.g. Ns- Nucleic acid sequences and X's -protein sequences.

By leaving out gaps, Tomkins is leaving out an entire group of mutations. Even worse, an ungapped analysis also produces lower identity scores than there should be. He is just wrong on this point. A gapped analysis is the only honest way of comparing the two genomes. I have shown why this is in previous posts

Loudmouth · Mar 19, 2013

mindlight said:
The idea of idel mutation e.g. insertions or deletions that determine phylogenetic direction assumes evolution.

No more so than the base substitions which Tomkins does include in his analysis. Why include base substitutions but exclude indels? The answer is obvious. Using an ungapped analysis gives Tomkins a lower score so that is what he goes with.

It may well be that species A has 4Gs at a locus where species B has 5Gs but that does not have to mean that B underwent a frameshift from the A like arrangement.

If you are comparing the genomes, then yes, you have to include this as an indel. That is how it works. Whether that got there by random mutagenesis or by design doesn't matter, it is still an indel and it needs to be treated as such.

This seems highly speculative to me and also contradicts the notion that each creature was created according to its kind. A creationist is not going to accept an evolutionary perspective and understanding of idels by default (nor is that a dishonest approach) and anyway as your links states it would be impossible to prove if species were gaining or losing sequences so the notion of phylogenetic direction is actually nonsensical.

It is no more speculative than evolution producing base differences. We have observed that humans are born with indel mutations, so we already know that nature produces these mutations, so why exclude them from the analysis?

Loudmouth · Mar 19, 2013

mindlight said:
I see similarities in the work of the creators hand and groups of creatures that share comon features and yet are all uniquely configured when considered holisitically.

But why a nested hierarchy? From a design standpoint, a nested hierarchy makes absolutely no sense. No human designer has his designs fall into a nested hierarchy. Even when humans design organisms they regularly violate the nested hierarchy?

So why a nested hierarchy? Why did God limit himself to a pattern of divergence and homology that evolution would produce?

The small differences in their coding here and there producing radically different results. The evidence supports this view better in my view

That's fine. This only demonstrates that a low mutation rate can produce the differences we see between species.

But no type to type evolutionary jumps have ever been observed and even by evolutionists own arguments would be unobservable because of the vast timespans in which these changes are supposed to have occurred. Microevolution does not prove macroevolution- it only proves the flexibilty of the Makers design to adapt to changing conditions.

Those mutations have been observed. They are the mutations that separate humans and chimps.

And-U-Say · Mar 19, 2013

mindlight said:
There is a certain pattern in the coding for life that reoccurs in slightly different ways in each creature according to its type and even further on occasions to specific creatures (as sometimes genes are on or off on an individual rather than collective basis)- that shows uniquely considered configuration. To me that reflects a common designer of an order of intelligence and with a work rate way beyond our own. It requires much more effort to try and cram the same facts (and we agree on these) into an all emcompassing theory of development that goes back billions of years which was initiated by accident and has acquired its own logic by trial and error.

What? No! I am talking about the re-use of structures. Why do a man's hand, a chimp's hand, a bat's wing, and a whale's flipper all have the identical bone structure? Re-using the same bone structure is sub optimal, something an infinite creator would not do. It is something the process of evolution would have to do. Again, this is evidence of evolution and completely refutes your creationism.

And-U-Say · Mar 19, 2013

mindlight said:
I think I find the computer code analogy better to the book analogy. Compare 2 pages of text on which a couple of words are different on one page rather than the other or a few are missing and I grant you the difference is .3% or whatever. But with the computer code analogy referencing different objects in a single line of code will produce an entirely different result to the point where I could say that even a single letters difference between 2 pages of otherwise identical code made the 2 sets of code incompatible and in that case I would be tempted to say there was a 0% alignment. Also one cannot just consider the single pages. Add up these differences over a whole book and you have an entirely different story. For example one story refers to Jim and another to sally. The story takes on a completely different meaning with just a single word changed with the implications of gender etc then thrown in.

IT IS NOT A COMPUTER CODE.

I am a Chemist, you need to know a lot more about Chemistry and its relation to Biology before you will understand why this analogy of yours is wrong.

mindlight · Mar 20, 2013

Loudmouth said:
By leaving out gaps, Tomkins is leaving out an entire group of mutations. Even worse, an ungapped analysis also produces lower identity scores than there should be. He is just wrong on this point. A gapped analysis is the only honest way of comparing the two genomes. I have shown why this is in previous posts

No for a number of reasons:

1) BLASTNs reliability on gapped analyses is not proven

2) These areas are excluded cause they are actually areas of low complexity e.g. Ns- Nucleic acid sequences and X's -protein sequences ,that are most often and uncontroversially very similar. So their inclusion distracts from the task of identifying more significant DNA differences.

mindlight · Mar 20, 2013

Loudmouth said:
It is no more speculative than evolution producing base differences. We have observed that humans are born with indel mutations, so we already know that nature produces these mutations, so why exclude them from the analysis?

We are all configured differently. Observed Micro-evolutionary changes do not prove theorised macro-level ones. The evolutionary theory is interfering with a proper analysis of the facts here by providing reasons to do things that distract from the task at hand and thereby distorting the results.

Tomk80 · Mar 20, 2013

mindlight said:
We are all configured differently. Observed Micro-evolutionary changes do not prove theorised macro-level ones. The evolutionary theory is interfering with a proper analysis of the facts here by providing reasons to do things that distract from the task at hand and thereby distorting the results.

This is in no way a response to the post of Loudmouth. Could you actually engage his argument please? Thanks!

Apes and humans have different designs

See in the dark

Senior Member

Senior Member

Learn whatever I can.

Senior Member

SCIENCE CAN TAKE A HIKE

See in the dark

His Supreme Holy Correctfulness

See in the dark

See in the dark

See in the dark

See in the dark

Contributor

Contributor

Contributor

Veteran

Veteran

See in the dark

See in the dark

Titleless

Similar threads

Privacy & Transparency

Privacy & Transparency