Vi bruger Cookies!     

         
 X     
intelligentdesign

Intelligent Design, Creationism and Evolution in Denmark and the rest of the world


The Mitochondrial Genome

For general background information on genetics, look here.

(Comments on Humans on selection and on the genome as such - scroll down)
To investigate the homology between two or more mitochondrial genomes, follow the instructions here.

Are there more than one kind of Rhino and Elephant?

According to creationism living organisms are found in separate 'Kinds'. Some often mentioned examples are the 'dog-kind' and the 'cat-kind'. This means that all cats are related, and all dogs are related, but there is no genetic relationship betwen dogs and cats.
This is in contrast to evolution. According to evolution, dogs and cats are families within the order Carnivora. All carnivores are more related to each other than they are to any other group of mammals.
Creationists more or less accept 'Kind' to be equal to 'Family' with some obvious exceptions. The family Hominidae (Great Apes) include Humans, Chimps, Gorillas and Orangutans. These species are of course not regarded as the same biblical 'Kind'.

According to creationism, after the Flood (some 4400 years ago), only one female from (most of) each 'kind' survived. The consequence of this is that within a kind, all differencies in the mitochondrial genome sequences, called mtDNA, are due to mutation of the mtDNA of this particular female. This phenomenon can be investigated fairly easily as mtDNA has been sequenced from a large number of species. I have looked into some of the slowest breeding land-mammals, the Rhinos and the Elephants, to see if the observations confirm the idea of biblical 'Kinds'.

There are five different species of Rhinoceros alive today:
White rhino, Ceratotherium simum (mtDNA, Genbank: NC_001808)
Black rhino, Diceros bicornis (mtDNA, Genbank: NC_012682)
Indian rhino, Rhinoceros unicornis (mtDNA, Genbank: NC_001779)
Javan rhino, Rhinoceros sondaicus (mtDNA, Genbank: NC_012683)
Sumatran rhino, Dicerorhinus sumatrensis (mtDNA, Genbank: NC_012684)

Of these, the largest difference is between the White and the Indian Rhino with almost 2000 different nucletides (nt). As there are a number of insertions/delitions involving more than one nt, the number of mutations (given these species has a common ancestor, and no selection has taken place) is probably closer to 1900.
Minimum generation time in Rhinos is about 6 years, so the maximum number of generations seperationg to species is about 1470 (2*4400/6), if one accept the creationist model. This means that 1900/1470 = 1.3 mutations pr generation have occurred. According to Jeanson (Ref: Here) the average mutation rate in humans is about 1 in 5 generations. This means that either there are more than one 'Rhino-kind', Rhinos has a mutation rate about 6.5 times as high as that of Humans, or the generation time of Rhinos was extremely short in the past. Neither of these solutions are plausible.

The same pattern is seen in Elephants.
African Elephant, Loxodonta africana (mtDNA, Genbank: NC_000934)
Indian Elephant, Elephas maximus (mtDNA, Genbank: NC_005129)
Number of differences (minus insertion/deletions): appr. 850
Minimum generation time: appr. 10 years
Maximum number of generations since the Flood: 880 (2 x 440 in 440 years)
Number of mutations pr. generation: 850/880 = 0,97 

If we include the Mastodon (Mammut americanum, GenBank: NC_009574) the situation becomes even worse for creationists.
When comparing African Elephant and Mastodon we find about 2000 nt differences.
Given the same generation time as in modern Elephants, the number of mutations pr. generations must have been about 2.3; 11.5 (2.3*5) times as much as in Humans.

If selection is included the situation of course gets even worse for the creationist model.

The conclusion is that neither elephants nor Rhinos can confirm the pattern expected from the creationist model.

Nuclear genes

See Ref 2.
Analyses of nuclear genes raises the same question of mutation rates.
Under the creationist model, after the Flood two indiduales had survived. These two individuals had up to four different alleles of each gene. The sequences of these original alleles is unknown. The analysis can therefor only make sense when comparing species where it can be shown that the gene have arisen from the same allele.

Within the Cat family, the IRBP-gene was analysed in the following species
(Systematic name, Common name, GenBank number).

Lynx lynx (Eurasian lynx) AY525038
Felis silvestris bieti (Chinese desert cat) AY525033
Prionailurus bengalensis (Leopard cat) AY525035
Otocolobus manul (Pallas's cat) AY525039
Uncia uncia (Snow leopard) AY525042
Neofelis nebulosa (Clouded leopard) AY525032
Panthera leo (Lion) AY525036
Panthera tigris (Tiger) AY525037
Panthera pardus (Leopard) AY525041

Results show at least four groups: Panthera/Uncia/Neofelis; Prionailurur/Otocolobus; Lynx; Felis.

In the following, each of these groups are interpreted as corresponding to each of the original four alleles.

Within the Panthera/Uncia/Neofelis group, the largest number of differences is 7.
The size of the analysed DNA-sequence is 1274.

The maximum number of generations is 8800 (2*4400)

Number of mutations pr. generation: 0.0008 (7/8800)
Number of mutations pr. generation pr. nt: 0.00000063
(0.0008/1274)
Number of mutations pr. 3 bill. nt (the size of the Human genome): appr. 1800

This should be compared to the number of mutations in Humans, appr. 60 pr. genome pr. generation (Jeanson in 'Searching for Adam'. p. 295. Look here)

This is under the assumption that the number of mutations is dependent on number of generations and not number of years. This assumption is relevant in mitochondrial DNA, but less relevant in nuclear genes. If we in stead calculate the number of mutations pr. year in Humans we reach about 3-4 mutations pr. year, dependent on generation time. Then the 1800 pr. year in cats becomes even more remarkable.

Either there are more than one Biblical 'Kind' of cats (which goes against the fact that species within the Panthera/Uncia/Neofelis group can breed with species outside the group), or mutation load is about 30 (250-300) times higher in cats than in humans. Neither are plausible solutions.

It is no help to assume that the alleles in the Panthera/Uncia/Neofelis group originated from more than one original allele. In this case at least one of the alleles must be present in one of the other groups (Prionailurus/Otocolobus, Lynx or Felis). The result of such a hypothesis is that that the number of acculated mutations are even bigger:

Panthera/Uncia/Neofelis vs Prionailurur/Otocolobus 10-13 differences.
Panthera/Uncia/Neofelis vs Lynx 16-19 differences.
Panthera/Uncia/Neofelis vs Felis 17-21 differences.

Another rescue device could be to postulate that these sequences originate form two or more loci, and therefore cannot be compared. But this is not a viable theory either. If all known sequences of the IRBP-gene are compared, the pattern of homology is incompatible with such a theory.

The same overall pattern was found when the TRSP-gene was analysed in the Canidae family. Though here it was difficult to identify four allele groups as suggested by the creationist model. Six groups seemed more plausible.


Homology between Human and Ape mtDNA

A somewhat laborious example of this is shown here:

The protein coding genes of mtDNA of a Human, Homo sapiens (GenBank: HQ260949) a Chimp, Pan troglodytes (JF727202) and a Gorilla, Gorilla gorilla (NC_011120) was compared. Both the DNA and protein sequences was compared.
The results was as follows (numbers are total number of differences)

DNA Homo Pan
Pan 1091  
Gorilla 1328 1281

 

Protein Homo Pan
Pan 161  
Gorilla 211 208

When Protein differences are subtracted from DNA differences, we get the synonymous differences (DNA-differences that do not change the amino acid in the protein sequence).

Synonymous Homo Pan
Pan 930  
Gorilla 1117 1073

The difference between Human/Chimp and Human/Gorilla is significant
(p < 0.002 - Poisson distribution with mean 1024)

The difference between Chimp/Human and Chimp/Gorilla is also significant
(p < 0.02 - Poisson distribution with mean 1002)

The difference between Gorilla/Human and Gorilla/Chimp is not significant
(p > 0.7 - Poisson distribution with mean 1095)

From a creationist point of view, it is hard to see why such a significant result, confirming the overall result from comparing total mtDNA, should be found.
Nathaniel Jeanson (Answers in Genesis, Ref: here) found that homology in animal mitochondrial proteins all in all followed major taxonomical clades above the family level (thought to be approximately equal to biblical 'kinds'). He suggested that the mitochondrial proteins served some kind of function relevant to anatomy or physiology.

This result go against this suggestion, as synonymous differences, per definition, has no influence on protein structure.

 

Selection within the Mitochondrial genome

Some of the questions addressed here is given a much more comprehensive treatment in Ref 1.

If mutations in the Mitochondrial genome (mtDNA) of humans were under strong selection, we would expect this to result in a non-random distribution of differences between two human mtDNAs.
This is exactly what can be found. I have compared two mtDNAs: KC345974 and FR695060 (Present day Human and Denisovan respectively), and recorded the differences according to where they appear.

The results are as follows:

Total mtDNA

Source Total size Differences %
tRNA   1489   18 1,21%
rRNA   2511   23 0.92%
Protein coding genes 11331 281 2.48%
D-loop   1120   56 5.00%
Total 16451 378

(intergenic spacers have been left out of the analysis)

The D-loop is known to be more prone to mutation, than the rest of the mtDNA. therefor calculation of an over-all mean is not that relevant. 

As can be seen, there is a factor 2,5 between the highest and the lowest rate (leaving out the D-loop), indicating strong selection at least in some areas of the genome.
(Significance of this result can be found below)

The protein coding genes

The number of amino acid substitutions within the protein coding genes were counted to a total of 41.
Synonymous mutations should on average be expected to amount to at least one in four of all mutations in protein coding genes. That 240 of 281 differences between the two genomes are synonymous shows a strong purifying selection working on non-synonymous mutations.
Under the simplest assumption: that mutations occur randomly in the protein coding genes, we can calculate that for every 240 synonymous mutations 720 non-synonymous must have occurred. Only 41 of these have survived selection.
This means that more than 9 in 10 of non-synonymous mutations have been lost in selection.
(synonymous and non-synonymous mutations have the same chance of being lost due to genetic drift)
This number is probably too high as mutations are known to be non-ransom, some combinations of nucleutides being more prone to mutaiton than others.

rRNA and tRNA

These genes are special in the sense that all nucleotides are potentially relevant to function.
How strong selection is in this area is hard to say, but as the analysis of the protein coding genes indicate, as much as 9 in 10 mutations might have been selected against. 

D-loop

What about the D-loop, the area with the highest frequency of differences? Are mutations randomly distributed in this area of the genome? Far from, as the next analyses show.
Had the mutations been randomly distributed, the number of nucleotides between two differences would follow a Poisson distribution.

The table should be read in the following way:
Obs: The number of nt between two subsequent differences.
Number: The number of areas of the given size.
Frequency: Number as rate of the total number of areas.
Exp. freq: The expected frequency, had the mutations been randomly distributed.
The last number was calculated using a Poisson distribution with mean 19.55, the mean size of the areas between two differences.

Obs Number Frequency Exp. freq.
0-10 30 0.56 0.014
11-20   9 0.17 0.583
21-50 12 0.22 0.401
51+   3
0.056 < 0.00001

 As is clear from the results, the differences are far from randomly distributed. Way too many are found close to, or far from, each other.
The longest areas between two differences are 70, 95 and 187 while according to the Poisson distribution, none should have been found longer than 40.
This indicates strong selection in some areas of the D-loop.
(the significance of this result is evaluated below)

All in all, there is little doubt that the mtDNA of Humans is under strong selection.

Of course, to further validate this, more sequence-pairs should be analyzed. But this is quite labor-intensive work, so I havn't had the time to do it. I invite anyone to do it.

Another related topic is the non-random use of synonymous codons in protein coding genes in the mtDNA.
I am currently looking into this question, which is more complicated than one might expect.
So far, it seems that selection and biased mutation rates work together to produce this phenomenon. The later papers emphasizing biased mutation rates over selection, but recognizing both.

Conclusion

All in all, it seems that an educated guess would be that a substancial fraction of mutations have been lost to selection. If that is so, the actual number of mutations would have been much higher than the 378 found in the table.
If Jeanson is right, that the overall mutation rate in mtDNA is about 1 in 5 generations, The 378 mutations corresponds to 1890 generations or 1890*20/2 = 18,900 years (5 generations per century in each line leading to Modern humans and Denisovans respectively). And this would be an absolute minimum.

Additional remarks

I made the same analysis of two mtDNAs from modern humans. The results confirmed both the results shown above, and the results found by others that the ratio of non-synonymous to synonymous mutations falls with growing total number of mutations (references in Ref. 1).
The interpretation of this phenomenon is that selection over the course of time removes more and more non-synonymous mutations, as some amino acid substitutions are only slightly harmful. Also rRNA and tRNA genes are thought to be under selection for slightly harmful mutations.


For those of you who have access to the scientific literature, here is a list of relevant recent papers:

  • Chen et al. Mutation and selection cause Codon Usage and Bias in Mitochondrial Genomes of Ribbon Worms (Nemertea), PLoS ONE 9(1) (2014):  e85631. doi:10.1371
  • Wei et al. Analysis of Codon usage bias of mitochondrial genome in Bombyx mori and its relation to evolution. BMC Evolutionary Biology 14 (2014):262-271
  • Uddin et al. Codon bias and genome expression of mitochondria ND2 gene in chordates. Bioimformation 11(8) (2015): 407-412
  • Uddin and Chakraborty Synonymous codon usage pattern in mitochondrial VYB gene in pisces, aves and mammals. Mitochondrial DN PartA (2017)

 

Further analyses of the results

I have conducted a Chi-square test of the results to further validate the non-random distribution of differences.
To do this the expected numbers should be calculated.
Expected numbers are the numbers that would be expected if the differences were randomly distributed. This calculated from the mean rate of differences and the total size of each area.

In the D-loop results, the two last cells had to be merged, as a Chi-square test is only reliable if all numbers are 5 or higher.

Total mtDNA


Number Exp
tRNA   18   34
rRNA   23   58
Protein coding genes 281 261
D-loop   56   26




p < 0.00001

D-Loop

Obs Number Exp
0-10 30   1
11-20   9 32
21+ 15 22




p < 0.00001

The p-value indicates the probability that differences are randomly distributed.

Note: The slight difference between the sum in the 'Number' and 'Exp' columns are due to rounding errors and do not influence the results, as the calculations were made on the actual numbers.


General conclusions
In all investigated cases reported above, the creation model falls short of its own predictions.
Both mtDNA and nuclear genes show that the predictionsfrom the creationist model is not met by the observations.

 

Ref 1:
Soares et al. Correcting for Purifying Selection: An Improved Human MitochondrialMolecular Clock. Am J Hum Genet 84: 740-759

Ref 2:
Yu et al. Phylogenetic relationships within mammalian order Carnivora indicated by sequences of two nuclear DNA genes Molecular Phylogenetics and Evolution 33: 694-705

Remark on the existence and nature of the mitochondrial genome

The Mitochondrial Genome (mtDNA) represent a problem for both creationists and evolutionary scientists.
Throughout, you should remember that the only known functions of mitochondrial genes are common to all organisms, and has nothing to do with anatomy. Look here for a creationist view on that question, and here for my response.

The problems from the creationist point of view are numerous. Here is a few.
1. Why is there a mitochondrial genome in the first place?
2. Why does phylogeny based on mtDNA mirror that based on nuclear DNA and anatomy, on all taxonomic levels?
3. Why is there separate genetic codes in different animal phylae?

1. I never saw anyone even try to address this problem. From an evolutionary point of view the explanation is clear and well understood. Mitochondria descended from a bacteria, that was once free living. Look here for more information.

2. From a creationist point of view you need to explain why mtDNA from animals that are placed in the same group due to anatomical features (like the Horse, the Rhinos and the Tapirs) have a greater homology between then than they have to other groups.
From an evolutionary point of view, this is exactly what you would expect due to a more recent common ancestor within the group than outside.

3. I never saw any creationist response to this question either.
From an evolutionary point of view, you can say that if you look for different genetic codes, mtDNA is the place to look, because of the very few translated genes.

To investigate the homology between mtDNA from different species, use this procedure:

1) Open www.ncbi.nlm.nih.gov.
2) In the 'All Databases' menu, select 'Nucleotide'.
3) In the input field write: 'complete mitochondrial genome' and the systematic (Latin) name of one of the organisms in question together with the family or order to which it belongs. E.g 'homo hominidae'. Press 'Enter'.
4) Now a number of hints come up. Select one of those that says something like 'mitochondrial DNA complete genome' it should be of about 16.000 - 17.000 bp.
5) Write down the Accession number (you don't need the number after the full stop). E.g. AP008824 (Homo sapiens).
6) Repeat point 3) and 4) with another species and family, e.g. 'pan hominidae' resulting e.g. in KM679417 (Pan troglodytes).
7) Scroll down and select 'BLAST' at the bottom of the page.
8) Select 'nucleotide blast'.
9) Select 'Align two or more sequences' and 'Somewhat similar sequences'.
10) Write one Accession number in the upper field and the other in the lower and press BLAST (you might have to down scroll a little).
11) After a few seconds the result come up. You need to scroll down a little to see that 99 % of the sequences match, and the match is 91 %.
A more detailed investigation and a little calculation gives the number 90,6 % homology when you compare the complete sequence.

If you want, many different sequences can be compared at the same time. Write one of them in the upper field and the rest in the lower. Be careful that all sequences are actually comparable. In this case it should be complete mitochondrial genomes. Partial genomes wouldn't do.
When several sequences are aligned, you can get a 'Distance tree of results' (a phylogenetic tree).
If you include outgroups, these should be placed in the lower field.

Specifically on Jeanson's work on the Human mtDNA

In a number of papers, Dr. Nathaniel Jeanson, together with coworkers, study the mtDNA in order to trace Human ancestry. Neanderthals and Denisovans (Homo sp. Altai) are usually considered fully human by creationists, and should therefore be included in such analysis, unless good reasons for not doing so can be given. In Jeanson 2015a, the author includes both Neanderthals, Denisovans and Homo heidelbergensis in the analysis of human ancestry, but dismisses the results. He calls these archaic humans ‘unusual’ and speculate that their divergence from modern humans are due to high mutation rates or unreliable sequencing. In Jeanson 2013, Jeanson 2015b, Jeanson and Lisle 2016, and Jeanson 2016, the authors do not include Neanderthals or Denisovans in the analyses, and make no comments as to why not. In Jeanson and Tomkins 2017 (p. 306), the authors state that the sequences of Neanderthal mtDNA are “… plagued with DNA contamination from microorganisms and modern human DNA …”. In addition to that, there is a lack of knowledge of generation time. They therefore conclude that Neanderthals should not be considered in analyses of human ancestry. All in all, three suggestions are offered to explain why Neanderthals are excluded from the analyses of human ancestry: High mutation rate, contamination, and no knowledge of generation time. If any of the two first were true, we should expect various Neanderthal mtDNA-sequences to be at least as different from each other as each of them are from mtDNA from modern humans. This is far from being the case. The largest number of differences between two mtDNAs from modern humans is about 120 nt. The differences between Neanderthal and modern humans lies between 180 and 240 nt. The largest number of differences between two Neanderthals is about 80 nt.

This pattern would be highly surprising if Neanderthal sequences were of poor quality, or under high mutation rates, as suggested.

As to the generation time argument, Neanderthals are considered fully human. Therefore, there is no reason to believe that they had generation times significantly different from modern humans. Generation times of humans living centuries or millennia ago is not known either. Using modern generation times in interpreting ancestry (as is done in Jeanson 2016) is no better than attributing any arbitrary generation time to Neanderthals. Default must be not to assume different generation times to humans living before the present, unless compelling arguments can be suggested.

Either the consequences of high mutations rates or contamination should be investigated in order to give a more solid argument to why this is a plausible explanation for the low homology between Neanderthals and modern humans, or the data from Neanderthals should be included in the analysis.

I haven’t found any additional arguments to why not to include Denisovans in the analysis. However, the above arguments for including Neanderthals are valid for Denisovans as well.

The conclusions of these papers would probably be highly influenced by the inclusion of Neanderthals and Denisovans. As mentioned, Neanderthals are about twice as different from modern Humans, as the most different modern Humans are from each other. Denisovans differ even more.

Conclusion

Neither poor sequence quality, high mutation rate nor unknown generation time are valid reasons for excluding archaic humans in the analysis of mtDNA. Dr. Jeanson should take these objections into consideration in his next paper on human ancestry based on mtDNA.

Calculation of homology

The homology mentioned is found by comparing the following GenBank numbers using BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi):
H. sapiens / H. sapiens: HM771221 and KJ669164 H. sapiens / Neanderthal: JN655834 and KC879692 (low). FJ770968 and KX198083 (high) Neanderthal / Neanderthal: KX198083 and KC879692

List of references

Jeanson N. 2013. Recent Functionally Diverse Origin for Mitochondrial Genes from ~2700 Metazoan Species. Answers Research Journal 6:467-501

Jeanson N. 2015a. Mitochondrial DNA Clocks Imply Linear Speciation Rates Within Kinds. Answers Research Journal 8:273-304

Jeanson N. 2015b. A Young Earth Creation Human Mitochondrial DNA “Clock”: Whole Mitochondrial Genome Mutation Rate Confirms D-Loop Results. Answers Research Journal 8:375-378

Jeanson N. and Lisle J. 2016. On the Origin of Eukaryotic Species’ Genotypic and Phenotypic Diversity: Genetic Clocks, Population Growth Curves, and Comparative Nuclear Genome Analyses Suggest Created Heterozygosity in Combination with Natural Processes as a Major Mechanism. Answers Research Journal 9:81-122

Jeanson N. 2016. On the Origin of Human Mitochondrial DNA Differences, New Generation Time Data Both Suggest a Unified Young-Earth Creation model and Challenges the Evolutionary Out-of-Africa Model. Answers Research Journal 9:123-130

Jeanson N. and Tomkins J. 2017. Genetics Confirms the Recent Supernatural Creation of Adam and Eve. Chapter 10 in ‘Searching for Adam’, Ed by Terry Mortenson. Master Books

 

Opdateret 12/06/2017