The role of positive selection in determining the molecular cause of species differences in disease
© Vamathevan et al; licensee BioMed Central Ltd. 2008
Received: 23 May 2008
Accepted: 06 October 2008
Published: 06 October 2008
Related species, such as humans and chimpanzees, often experience the same disease with varying degrees of pathology, as seen in the cases of Alzheimer's disease, or differing symptomatology as in AIDS. Furthermore, certain diseases such as schizophrenia, epithelial cancers and autoimmune disorders are far more frequent in humans than in other species for reasons not associated with lifestyle. Genes that have undergone positive selection during species evolution are indicative of functional adaptations that drive species differences. Thus we investigate whether biomedical disease differences between species can be attributed to positively selected genes.
We identified genes that putatively underwent positive selection during the evolution of humans and four mammals which are often used to model human diseases (mouse, rat, chimpanzee and dog). We show that genes predicted to have been subject to positive selection pressure during human evolution are implicated in diseases such as epithelial cancers, schizophrenia, autoimmune diseases and Alzheimer's disease, all of which differ in prevalence and symptomatology between humans and their mammalian relatives.
In agreement with previous studies, the chimpanzee lineage was found to have more genes under positive selection than any of the other lineages. In addition, we found new evidence to support the hypothesis that genes that have undergone positive selection tend to interact with each other. This is the first such evidence to be detected widely among mammalian genes and may be important in identifying molecular pathways causative of species differences.
Our dataset of genes predicted to have been subject to positive selection in five species serves as an informative resource that can be consulted prior to selecting appropriate animal models during drug target validation. We conclude that studying the evolution of functional and biomedical disease differences between species is an important way to gain insight into their molecular causes and may provide a method to predict when animal models do not mirror human biology.
Much scientific and medical progress has depended on experimental findings in model organisms being extrapolated to humans. However, even closely related species such as humans and chimpanzees, often experience the same medical condition with varying symptomatology, as seen in cases of Alzheimer's disease or AIDS, or with varying prevalence, for example, autoimmune diseases, epithelial cancers and schizophrenia [1, 2].
Comparison of disease prevalence and symptomatology across species is complicated by the fact that modern human lifestyles, very far from the conditions of early human evolution, may reveal susceptibilities to disease that were not evident in the early history of the human species . However, there are observed biomedical differences between humans and other animals that cannot be wholly explained by lifestyle [1, 2].
Genetic disease can occur as a by-product of an adaptation which confers a large selective advantage . For instance, the seemingly human-specific disease of schizophrenia  and the greater human susceptibility to Alzheimer's disease compared with primates  may be a by-product of the human specialisation for higher cognitive function . Besides Alzheimer's disease and schizophrenia, many other diseases also differ in frequency and symptomatology between humans and other mammals. Olsen and Varki  and Varki and Altheide  list some of these diseases with the emphasis on non-human primates, indicating that for these diseases chimpanzees are not good models despite their close evolutionary relationship with humans. Genes that have been subject to adaptive evolution since the divergence of humans and other primates may be involved in this variation of phenotype and be key to understanding the disease state. Thus, comparative evolutionary genomics can offer insights into these disease mechanisms by correlating molecular differences that arose during species evolution with phenotypic differences in diseases between species; hence elucidating disease-causative genes and pathways.
Direct comparisons of human genomic and transcriptomic information to that of other species reveals three major types of molecular genetic changes which have contributed to species differences. The most obvious mode is the presence or absence of genes in different species, including gene duplication and gene inactivation. Much attention has been paid to genes that are unique to humans or lost in the human lineage [1, 2, 8, 9]. However these probably represent the 'tip of the iceberg' of human genomic differences compared to other species. The second class of molecular genetic changes constitutes of nucleotide substitutions that may cause functional changes in both protein coding and non-coding RNAs. The third category of molecular changes consists of variation in the levels of gene expression between species and in the mechanisms regulating gene expression [8, 10].
In this study we investigate the second type of molecular differences, and focus on coding changes in protein-coding orthologous genes. An estimated 70% to 80% of orthologous protein sequences are distinct between humans and chimpanzees [8, 9, 11]. However, a substantial proportion of differences may have no functional impact on human-specific diseases. Positive selection analyses can determine which nucleotide changes contribute to biological differences between species. This follows from the premise that the action of positive selection pressure in orthologous genes during evolution is often associated with sub- or neofunctionalisation of genes . Determining such genes on the human lineage is thus a rational and promising way to reveal the molecular changes implicated in human-specific disease.
In contrast to previous studies [13–17] which focused on human evolution, the objective of this study was to determine genes which have undergone adaptive evolution in both humans and animal models. We have analyzed alignments of 3079 orthologous genes from human, chimpanzee, mouse, rat and dog to detect signals of positive selection. These species were chosen as they are common models of human disease in medical research and high-quality genomic sequences were available.
Our initial dataset was aggressively filtered to eliminate paralogous alignments, spurious annotations, pseudogenes in one or more species, and poor exon prediction. Hence only quintets for which we could assign orthology with high confidence were used in our analysis for positive selection. Due to this strict screening it must be noted that our orthologue dataset may contain a bias towards orthologues of high levels of conservation, thereby underestimating the number of positively selected genes and underestimating the average levels of divergence. The direction and strength of selection is measured by ω, the nonsynonymous to synonymous substitution rate ratio (dN/dS = ω), with ω <1, = 1, and > 1 indicating purifying selection, neutral evolution, and positive selection, respectively. The branch-site model, which tests for positive selection that affects a small number of sites along pre-specified lineages [18–20] was used to test all extant and ancestral lineages for evidence of positive selection. The branch-site model has been shown to be more powerful and more conservative than methods that test positive selection on a given lineage or on a subset of sites . We identified genes predicted to have changed function during mammalian evolution and relate our findings to the diseases known to show biomedical differences between humans and model organisms. These genes may be causative of the phenotypic disease differences between species and are promising targets for therapeutic intervention. This approach is of interest to drug development as detection of positive selection in a drug target or members of a disease pathway may cause animal models to be non-predictive of human biology and explain some observed biomedical differences between species .
We found the chimpanzee lineage had many more genes under positive selection than any of the other lineages and three times more than the number of genes in the human lineage. We present evidence to argue against the possibility that this result is due to artefacts introduced by genome sequence coverage, gene sample selection or algorithmic sensitivity to errors in sequence data or alignments. Instead, we conclude that the elevated number of chimpanzee positively selected genes is a true reflection of evolutionary history and is most likely due to positive selection being more effective in the large population sizes chimpanzees have had in the past or possibly remarkable adaptation in the chimpanzee lineage.
As demonstrated in the yeast protein interaction network, evolutionary rate is thought to be correlated with protein connectivity [22–24]. Hence, genes under positive selection are generally believed to be less promiscuous, that is, they interact with fewer genes compared to genes under neutral evolution or negative selection. This may be because promiscuous genes are subject to functional constraints due to their pivotal or multiple roles in biological pathways. However, others analyzing the same data claim that the results are inconclusive [25, 26]. We investigate whether genes under adaptive evolution interact with fewer genes compared to genes not under positive selection but did not see a significant difference. However, we also investigated the hypothesis that a gene under adaptive evolution would drive complementary divergence of genes encoding interacting proteins. The most common examples of this co-evolution of interacting genes are receptor-ligand couples that co-evolve to maintain or improve binding affinity and/or specificity. Examples of such genes include the prolactin (PRL) gene and its receptor (prolactin receptor, PRLR) in mammals , primate killer cell immunoglobulin-like receptors (KIRs) that co-evolved with MHC class I molecules  and red and green visual pigment genes . Here we present evidence that positively selected genes are significantly more likely to interact with other positively selected genes than genes evolving under neutral evolution or purifying selection.
Detection of genes under positive selection
Number of genes under for positive selection in the seven lineages and number of positive genes in OMIM
There were several genes that showed signatures of selection in multiple lineages. We found 17 PSGs along both human and chimpanzee lineages, 8 PSGs along both mouse and rat lineages and 8 PSGs along the hominid and murid lineages. These numbers are significantly greater than we would expect by chance (e.g. there were more genes positively selected in both the human and chimpanzee lineage than would be expected by chance; p < 6.864e-10, Fisher's test; see Additional File 2, Table 1). Detailed analyses of the genes that overlap between lineages can be found in Additional File 2, 'Genes under selection in adjacent lineages' and Additional File 3.
Elevated numbers of positively selected genes were detected on the chimpanzee lineage
We found 162 PSGs along the chimpanzee lineage which was three times more than the 54 PSGs detected on the human lineage. This finding was in agreement with other reports of high number of genes that underwent positive selection during chimpanzee evolution [16, 31]. Bakewell et al.  (using a wholly different methodology to this study) identified 21 positive chimpanzee genes and 2 positive human genes from an initial data set of 13,888 genes. Elevated numbers of PSGs along the chimpanzee lineage were also found by Arbiza et al.  a more similar approach who identified 1.12% of genes under positive selection in the human genome and 5.96% in the chimpanzee genome, which is in close accordance with 1.75% (human) and 5.26% (chimpanzee) obtained here.
Functional processes affected by positive selection
OMIM is enriched for positively selected genes
In order to determine if our dataset of PSGs was significantly enhanced for disease genes, we examined genes that were associated with human diseases as defined by the OMIM, Online Mendelian Inheritance in Man database . Of the 3079 genes used in our analysis, 469 genes (15.2 %) were associated with a disease term in OMIM. Of the 511 PSGs from all seven lineages, 99 genes (19.4 %) were associated with a disease term in OMIM (Table 1). A test based on the binomial distribution showed that there is a significant link between PSGs and disease (p = 0.0067). While PSGs along the murid lineage were significantly over-represented in OMIM (p = 0.0087), PSGs along the human, chimp or hominid lineages did not display any over-representation (significance cut-off p = 0.05).
No correlation of PSGs and recent selection in human populations
We did not see any evidence of a relationship between a gene being positively selected within human populations and in our mammalian species. In fact, there seems to be a trend that suggests that genes are less likely to have been subject to positive selection along the hominid branch if they were under selection in recent human history. The number of human PSGs was compared with genes shown to be under positive selection pressure within human populations . This is evident in the lower proportion of genes that were both under recent positive selection and positively selected along the human branch (0.03%) compared to the proportion of genes under positive selection along the hominid branch alone (1.8%).
PSGs on all lineages show evidence of co-evolution
To test if PSGs or proteins encoded by PSGs interact with fewer genes or proteins compared to genes that are not under positive selection, we queried a meta-database of biological interactions (see Methods, ) with the list of all PSGs. For the 511 PSGs along all lineages, 155 (30%) did not have any annotated interactions with any other proteins and the median number of interactions was 5. For the 2568 genes in the test set with no evidence of positive selection, 783 (31%) did not have any interactors and the median number of interactors was also 5. Therefore PSGs do not have a lower median number of interactors than genes not under positive selection in the test set (p = 0.815; two-tailed Wilcoxon rank sum test), which suggests that number of interactors is not a determinant for PSGs.
Interacting clusters formed between PSGs on each lineage
pvalue for cluster size
given previous clusters
pvalue for cluster given
number of interactions per gene **
PCSK5, BMP4, PHOX2A
LHB, OTX1, JUB
CFP, TAL1, SERPINB1, MMP12, PRF1, BCL2, HRG, ITGA5, COMP
CD79A, HCLS1, LCP2
SNTA1, DAG1, MUSK
CCL19, CD86, MADCAM1
TLR5, CD86, PTGIR
SCNN1G, SPTA1, HECW1
CDKN2D, TRIM21, CDKN1B, CAST, ICAM1, CFD, ITGB2, C3
KCNA4, ACTN2, PIK3R5
We also tested each cluster to determine whether the size of the cluster is more than would be expected by chance given the number of interactors for each individual gene in the cluster. All 28 clusters were found to be significant (p < 0.05 by permutation test) (Table 2), therefore there is a significant phenomenon of PSGs interacting with other PSGs. To confirm this observation, a further analysis was performed on the genes that interact with the beta 2 integrin gene (ITGB2) which showed evidence of positive selection along the rat (p < 0.001) and murid (p < 0.05) lineages. Three of its four known interacting alpha subunits  also showed positive selection either on the murid branch (ITGAL, p < 0.01; ITGAX, p < 0.05) or on the mouse branch (ITGAD, p < 0.001).
The functional categories enriched for PSGs in this study were found to closely correlate with those detected in previous genome scans . The consensus is compelling given the different techniques used in each study and the risk of false positives inherent in large-scale studies. It is interesting to note that among the five species analyzed, protein families with distinct functions could be identified as evolving under positive selection for each species. Molecular changes in these genes are potentially responsible for driving the species-specific differences.
Hypotheses to explain the high number of PSGs on the chimpanzee lineage
The high number of PSGs along the chimpanzee lineage cannot be explained by the incorrect calling of orthologues or alignment quality, as we employed conservative filters during the orthologue calling procedure and manually checked all the PSG alignments. We also checked the underlying genomic quality values for the chimpanzee PSGs and only 1 sequence had quality values less than Q20 (error rate of 0.01) among the sites predicted to be under positive selection and hence the high number of PSGs is not due to poor genomic sequence quality. However, we acknowledge that the chimpanzee genome sequence is unfinished and will contain errors and rare polymorphisms, as exemplified by its occasional mismatches to mRNA and gene prediction sequences (such as those provided by RefSeq). In this study, we have tried to minimise the effect of sequence error by preferentially using validated gene sequences when available and high quality genome sequence when not. Nevertheless, we cannot exclude sequence error as a factor in our results. Therefore, we also checked that taxon sampling did not affect the number of PSGs on other lineages and hence ensured that quality issues from one species did not affect the signals for positive selection on other lineages (see Additional File 2 'Taxon sampling does not affect detection of positive selection' and Additional File 4). Additionally, comparison of 11 of the extremely divergent chimpanzee sequences to their orthologues in other primates (marmoset, macaque and orang-utan) (see Additional File 2 'Chimpanzee PSGs are lineage-specific') showed that the amino acid differences observed in the 11 chimpanzee sequences are specific to the chimpanzee, with the other primate sequences having the same state as the human sequence.
One likely explanation for the high number of PSGs in the chimpanzee lineage could be the reported high polymorphism in the individual chimpanzee sequenced (heterozygosity rate of 9.5 × 10-4 ). This rate is slightly higher than what was seen among West African chimpanzees (8.0 × 10-4 ) which have similar diversity levels to that seen in human populations . Population size is another possible explanation as positive selection may have had a reduced efficacy in humans than in chimpanzees due to the larger long-term population size of chimpanzees compared to humans indicated by reduced nucleotide diversity and elevated polymorphism among chimpanzee sequences .
PSGs implicated in diseases with biomedical differences between mammals
Overall, we observed that PSGs were over-represented among genes found in OMIM. Yet in contrast to the findings of Clark et al. , PSGs along the human lineage were not seen to display any over-representation in OMIM. Our findings, however, were consistent with other recent studies that found no significant associations  or only marginal associations  between human PSGs and human diseases. The OMIM database is the most complete freely available source of disease associated genes available but does include genes associated with non-pathological conditions such as hair colour; hence noise from such data might lead to non-significant results during statistical tests. Tests for enrichment of PSGs within more precise collections of disease genes may yield different results.
Examination of individual PSGs along the human and hominid lineages, revealed genes implicated in diseases that show biomedical differences between mammals. Below we illustrate how some of the human and hominid PSGs identified in our study are linked to medical conditions described as being more prevalent or having increasing severity in humans compared to apes [1, 2].
Human epithelial cancers are thought to be the cause of over 20% of deaths in modern human populations whereas among non-human primates, the rates are as low as 2–4% . Although this may be partly attributed to carcinogenic factors in the lifestyles of modern humans and differences in life expectancy, there are many intriguing lines of evidence to suggest that another overwhelming factor is the presence of susceptibility genes in human [8, 42–47].
Among the human lineage PSGs detected here a number of genes have been implicated in the development of epithelial cancers:
MC1R(melanocortin-1 receptor) modulates the quantity and type of melanin synthesised in melanocytes. Mutations in this gene have been associated with melanomas . An allele of this gene associated with pale skin colour and red hair, was recently located in the Neanderthal sequence  which suggests that this gene was also under recent selection in human evolution. Functional changes in the human MC1R gene which causes a change in skin colour could lead to an increased susceptibility to ultra-violet radiation and hence higher levels of melanoma in humans.
The G-protein coupled receptor EDNRB(endothelin type-B receptor) and its physiological ligand, endothelin 3, are thought to play key roles in the development of melanocytes and other neural crest lineages . EDNRB promotes early expansion and migration of melanocyte precursors and delays their differentiation. EDNRB is greatly enhanced during the transformation of normal melanocytes to melanoma cells where it is thought to play a role in the associated loss of differentiation seen in melanoma cells .
The presence of the ALPPL2gene product, an alkaline phosphatase isoenzyme, has been shown to increase the potential of premeiotic male germ cells to malignant transformation. Increased promoter activity of this gene was seen in the process of tumour progression. ALPPL2 has now been confirmed as a marker for testicular germ cell tumours .
GIPC2mRNAs are expressed in cells derived from a diffuse-type of gastric cancer, and also shows increased expression in several cases of primary gastric cancer . The PDZ domain of the GIPC2 protein interacts with several genes that are involved in modulation of growth factor signalling and cell adhesion (e.g. FZD3, IGF-1 and NTRK1). Thus GIPC2 may play key roles in carcinogenesis and embryogenesis.
In the hominid lineage, several PSGs have also been implicated in epithelial cancer development suggesting differences in cancer disease processes between hominids and other mammals:
MSH2is a DNA mismatch-repair gene that was identified as a common locus in which germline mutations cause hereditary nonpolyposis colon cancer (HNPCC) . As deficiencies in any DNA repair gene would potentially increase cancer risk, this group of genes is of interest in investigation of species differences in cancer prevalence. We found that genes which are involved in DNA repair and nucleotide metabolism were over-represented for PSGs along the chimpanzee and human lineages respectively (Figure 2). Enrichment of PSGs within the nucleotide metabolism category has also been reported previously .
The ABCC11[ABC-binding cassette, subfamily C, member 11] gene product is highly expressed in breast cancer compared to normal tissue. ABCC11 is regulated by ERα, which mediates the tumour promoting effects of estrogens in breast cancer .
Ataxia and Migraine
The calcium channel gene, CACNA1A, was found to be under positive selection along the human lineage. In humans, mutations in CACNA1A are associated with channelopathies, such as spinocerebellar ataxia 6 and episodic ataxia type 2  as well as with more prevalent conditions such as familial hemiplegic migraine, dystonia, epilepsy, myasthenia and even intermittent coma . It is possible that the trafficking or signal modulation of CACNA1A differs between humans and other mammals as a result of adaptation of the central nervous system, which could result in humans being more prone to these neurological disorders. The benefits of enhanced CNS excitability may outweigh the risk of severe headache and disability, the symptoms of migraines . It could also be an artefact of design constraints in the brain resulting from imperfect interconnections between older and more recently evolved brain structures .
A gene implicated in Alzheimer's disease [59, 60], APOE, was under positive selection along the hominid lineage. Selection for functional changes of the APOE gene in the hominid lineage could be related to either its role in neurological development or in lipid metabolism. Of the eight amino acids found to be under positive selection in this study, four are present in the lipid-binding carboxyl terminus.
The suggestion that there are species differences in Alzheimer's disease between humans and other mammalian species comes from the lack of pathological lesions including the neurofibrillary tangles associated with human Alzheimer's disease being observed in the brains of elderly chimpanzees [6, 61] or elephants . Also, transgenic mouse models of Alzheimer's disease that presented β-amyloid neuropathology do not exhibit the cognitive decline at the first appearance of amyloid plaques seen in humans . Finally and intriguingly, mammals other than humans seem to have just one allelic form of APOE, the E4 allele [60, 64], the same form in humans predisposes carriers to a much higher risk of Alzheimer's disease .
We hypothesise that the positive selection pressure acting on APOE during hominid evolution changed the role of APOE in neurological development, presumably in concert with the expansion of cognitive ability. However, alternative studies have suggested that the major evolutionary events associated with cognition have occurred much earlier . A consequence of increased cognitive ability maybe increased susceptibility to dementing diseases such as Alzheimer's disease  but as the onset of these diseases is past reproductive age, these diseases would be overlooked by natural selection. The other possibility is that dietary pressures influenced the evolution of APOE in mammals, with species adapting to diets with differential levels of lipids and so favouring different forms of APOE .
Neurological studies have shown that brain areas differentially dysregulated in schizophrenia are also subject to the most evolutionary change in the human lineage . A number of PSGs along the human lineage are associated with schizophrenia:
SNPs in the gene PIK3C2G[phosphoinositide-3-kinase] have been shown to be associated with schizophrenia recently . This gene is related to the phosphoinositide pathway, and thus is a probable candidate for schizophrenia and bipolar disorder .
Another candidate for chronic schizophrenia is the Q399 allele of the XRCC1protein, which plays a role in base excision repair . The pathophysiology of schizophrenia is associated with an increased susceptibility to apoptosis. Mutations in XRCC1 may cause DNA damage which if detected cause apoptosis regulators to arrest cell cycle progression.
Other cognitive disorders
Also subject to positive selection along the human lineage was the gene GFRA3, a receptor for artemin and a member of the glial cell line-derived neurotrophic factor (GDNF) family of ligands. This gene acts as a signalling factor regulating the development and maintenance of many sympathetic neuronal populations . In particular, along with other GDNF family members, artemin plays a role in synaptic plasticity, a mechanism thought to be central to memory . Deficiencies in GFRA3 would be expected to cause cognitive impairment making it a candidate gene for cognitive disorders.
Autoimmune diseases are rare in non-human primates whereas they are relatively common in humans . CENP-Bis one of three centromere DNA binding proteins that are present in centromere heterochromatin throughout the cell cycle. Autoantibodies to these proteins are often seen in patients with autoimmune diseases, such as limited systemic sclerosis, systemic lupus erythematosus, and rheumatoid arthritis . The positive selection pressure acting on this gene during human evolution is consistent with experimental results that antigenic specificity in the C-terminus of CENP-B is species-specific .
Positive selection of regulatory genes
Selection events on coding sequences may also have effects on gene expression regulation. One transcription factor that showed signs of positive selection along the human lineage was HIVEP3(immunodeficiency virus type I enhancer binding protein 3). This gene belongs to a family of zinc-finger proteins whose functions include activating HIV gene expression by binding to the NF-kappaB motif of the HIV-1 long terminal repeat . It is commonly known that HIV infection in chimpanzees does not progress to the level of medical complexity that is seen in human AIDS . In chimpanzees the virus lives in a benign relationship within the immune system whereas in humans it infects and destroys helper T-cells. Functional changes in transcription factors such as HIVEP3 between humans and chimpanzees could explain the observed differences in HIV disease progression.
Regulatory elements of gene expression also showed evidence of positive selection along the human lineage. One is the MOV10gene (Moloney leukaemia virus 10, homolog), an RNA helicase contained in a multiprotein complex along with proteins of the 60S ribosome subunit. MOV10 is associated with human RISC (RNA-induced silencing complex) . RNA silencing or interference (RNAi) has been recently described as an important therapeutic application for modulating gene expression at the transcript level or for silencing disease-causing genes [79, 80]. Any functional changes in the MOV10 gene due to selection may affect transcriptional control of multiple genes and would therefore prompt widespread differences among species.
We conclude that comparative evolutionary genomics has an important contribution to make to the study of mammalian disease, enabling identification of candidate genes for further in vivo investigation. Researchers traditionally see the biomedical differences between humans and model organisms as an obstacle to progress. However, we propose these differences also provide an opportunity to dissect the molecular causes of disease. To take advantage of this opportunity, we need powerful computational evolutionary algorithms (such as used in this study) and a robust approach to utilise the ever-expanding genomic sequence data. Two major challenges inherent to this approach are: firstly, sequence errors are likely to increase the false positive rates in identifying cases of positive selection pressure and secondly, to fully utilize this information requires detailed accounts of the physiological differences in disease occurrence and symptomatology between species which are currently sparse.
Understanding the evolutionary history of disease genes can also significantly impact the choice of pre-clinical animal models in the drug discovery process . The success rates in pharmaceutical pipelines remains low, one reason being the difficulty in successfully translating safety and efficacy studies from animal models to humans. Pre-clinical studies assume that drug targets in the experimental species and in humans are functionally equivalent, which is not always the case . In particular, animal models of neurodegenerative diseases have been shown to lack predictive validity in humans . Studies of selection pressure during gene evolution can provide valuable information for the choice of animal models for drug target validation. Our results of PSGs in the five mammalian species serve as an informative resource that can be consulted prior to selecting appropriate animal models during drug target validation in the pharmaceutical industry.
Positive selection pressure would be expected to act not just on one gene at a time but on pathways of genes. We found that genes that were subject to positive selection along the same lineage were significantly more likely to interact with each other than with genes not under positive selection, the first evidence for co-evolution of genes as a widespread phenomenon in mammals. We suggest that the high level of connectivity between PSGs is caused by compensatory change of a protein's interaction partners when a protein undergoes change in response to selection.
We observe many chimpanzee genes which have been subject to positive selection during the evolution of their anthropoid ancestor. Since medical research and the vast majority of biological research have been focussed on discovering more about human biology, we know a lot less about chimpanzee-specific characteristics. The number of PSGs on the chimpanzee lineage would suggest that these chimpanzee adaptations are at least as striking as our much-vaunted human-specificities.
We analysed all Entrez human genes (accessed in September 2006) that were annotated as protein coding and had a confirmed mRNA sequence. The longest open reading frame associated with each gene was included in the starting set. Curated mRNA sequences from the RefSeq NCBI database and genomic sequences for the four model organisms (chimpanzee, mouse, rat and dog) and chicken (outgroup) were extracted from GenBank (accessed in September 2006).
The orthologue detection pipeline used reciprocal tBlastX searches  between the human and model organism sequence databases. If the highest scoring non-human species sequence was genomic, indicating an mRNA sequence was not available for this gene in this species, it was processed via GeneWise  to identify a predicted gene structure and remove introns, using the human peptide as template. The resulting cDNA sequence was then used as a query in the reciprocal tBlastX search against the human database. Highest scoring mRNA sequences were submitted to the reciprocal tBlastX search without modification.
Reciprocal best hits between the human gene and the model organism gene were marked as the orthologue pair for that human transcript query on the condition that the log of the p value from the best hit of the human mRNA sequence against the model organism database was higher than 95% of the log of the p value of the best hit from the reciprocal step.
Incomplete genome sequencing will also contribute to error in orthologue calling. Reciprocal blasting is invalidated as a method for calling orthologues in these circumstances as the absence of the true orthologue would cause a more divergent paralogue to be the top hit. To address this problem we added a cut-off, which required the p value of the putative orthologue for that species to be less than that of the chicken orthologue for that gene. The chicken was chosen because it was the closest relative to mammals for which a complete draft genome sequence was available at sufficient coverage . For the 262 human genes with no chicken orthologue, those predicted by reciprocal BLAST alone were analysed but these genes were flagged as potential problems.
Detecting genes affected by positive selection
The resulting sets of 5 orthologous sequences were translated and aligned using Muscle , then converted to corresponding nucleotide alignments. All alignments were then corrected for frameshifts in the sequences from the model organisms relative to human. Unrooted tree files for each alignment were created using a standard mammalian species tree  ((human, chimpanzee), (mouse, rat), dog) (Figure 1). Initially, data sets were analyzed using the M0 (one-ratio) model implemented in the codeml program from the PAML package . The M0 model assumes constant ω ratio for all branches in the tree and among all codon sites in the gene . Two runs of the M0 model were performed on each alignment to check that values for log-likelihood, κ and branch lengths were consistent between the two runs. Runs that were not consistent were rerun until the values converged. In the subsequent analyses using the branch-site model, the branch lengths and the transition/transversion rate ratio κ were fixed to their estimates under the M0 model. This strategy reduces the computation time as the number of parameters to be estimated is reduced.
To infer the lineage specific evolution of genes, the branch-site model [18, 19] was used to test for positive selection. We tested each of the seven branches on the species phylogeny, treating each in turn as the foreground branch. Results prior to multiple hypothesis correction should not be used for subsequent analysis as the family-wise error rate is unacceptably high . Here we report results following a Bonferroni correction for multiple testing which is known to be conservative and hence, prediction of positive selection is particularly robust. The corollary of such a strict approach is the potential generation of false negatives. The alternative branch-site model has four codon site categories, the first two for sites evolving under purifying selection and neutral selection on all the lineages and the additional two for sites under positive selection on the foreground branch. The null model restricts sites on the foreground lineage to be undergoing neutral evolution. Each branch-site model was run at least three times to ensure convergence of log-likelihood values at or within 0.001. Runs that did not converge with additional runs indicated problems with the data and reported as such.
When the data from the automated procedures was examined closely, it was noted that some alignments had areas of ambiguous alignment or areas where sequences did not appear orthologous. Areas of non-orthology could result from incomplete gene predictions due to gaps in the genomic sequence or absent or variant exons. Therefore the data were subjected to further manual corrections detailed below:
1. To correct for regions of low similarity, all alignments were scanned to mask out parts of a sequence where > 3 consecutive codons were different to the other sequences in the alignment and where these codons were flanked by gaps on one or both sides. Sequences that also contained frameshifts relative to the human sequence were corrected.
2. After re-running PAML on the entire dataset, we manually examined the alignments of all significant results (p < 0.05). The result was discarded if the gene sequence belonging to the lineage that was identified as being under positive selection had a frameshift or was ambiguously aligned.
Analysis of interaction data
A network consisting of protein-protein interactions such as binding and phosphorylation, transcriptional control and post-translational modification was used to search if genes under positive selection interact together. Interaction data in the network was licensed from several commercial vendors including Ingenuity , Jubilant , GeneGO , NetPro  and HPRD . All of the information from these databases is based on manual curation of literature. In addition, high-quality, automatically extracted interactions licensed from the PRIME database  were also included in the network. Interactions associated with transcriptional regulation were obtained from experimental validation protein-DNA binding relationships licensed from the TransFac  and TRRD  databases. No distinction is made between DNA, RNA and protein for a particular gene, and all three are represented as a single node in the network. Searches of gene lists that resulted in a biological sub-network were conducted and scored as in .
Positively Selected Gene.
We would like to thank Fabrizio Caldara for his help with the disease ontologies and Roberto Alvarez for his help with sequence databases. We are also grateful to three anonymous reviewers for their thorough and constructive comments that helped to improve the manuscript. This study was supported by a grant from the Biotechnological and Biological Sciences Research Council (BBSRC) to ZY, and an MRC Bioinformatics Fellowship to RDE.
- Olson MV, Varki A: Sequencing the chimpanzee genome: insights into human evolution and disease. Nat Rev Genet. 2003, 4 (1): 20-28. 10.1038/nrg981.View ArticlePubMedGoogle Scholar
- Varki A, Altheide TK: Comparing the human and chimpanzee genomes: searching for needles in a haystack. Genome Res. 2005, 15 (12): 1746-1758. 10.1101/gr.3737405.View ArticlePubMedGoogle Scholar
- Young JH, Chang YP, Kim JD, Chretien JP, Klag MJ, Levine MA, Ruff CB, Wang NY, Chakravarti A: Differential susceptibility to hypertension is due to selection during the out-of-Africa expansion. PLoS Genet. 2005, 1 (6): e82-10.1371/journal.pgen.0010082.PubMed CentralView ArticlePubMedGoogle Scholar
- Nesse RM, Williams GC: Why we get sick: the new science of Darwinian medicine. 1995, New York: Times BooksGoogle Scholar
- Crespi B, Summers K, Dorus S: Adaptive evolution of genes underlying schizophrenia. Proc Biol Sci. 2007, 274 (1627): 2801-2810. 10.1098/rspb.2007.0876.PubMed CentralView ArticlePubMedGoogle Scholar
- Gearing M, Rebeck GW, Hyman BT, Tigges J, Mirra SS: Neuropathology and apolipoprotein E profile of aged chimpanzees: implications for Alzheimer disease. Proc Natl Acad Sci USA. 1994, 91 (20): 9382-9386. 10.1073/pnas.91.20.9382.PubMed CentralView ArticlePubMedGoogle Scholar
- Keller MC, Miller G: Resolving the paradox of common, harmful, heritable mental disorders: which evolutionary genetic models work best?. Behav Brain Sci. 2006, 29 (4): 385-404. discussion 405-352PubMedGoogle Scholar
- Kehrer-Sawatzki H, Cooper DN: Understanding the recent evolution of the human genome: insights from human-chimpanzee genome comparisons. Hum Mutat. 2007, 28 (2): 99-130. 10.1002/humu.20420.View ArticlePubMedGoogle Scholar
- Chimpanzee SaAC: Initial sequence of the chimpanzee genome and comparison with the human genome. Nature. 2005, 437 (7055): 69-87. 10.1038/nature04072.View ArticleGoogle Scholar
- Gilad Y, Oshlack A, Smyth GK, Speed TP, White KP: Expression profiling in primates reveals a rapid evolution of human transcription factors. Nature. 2006, 440 (7081): 242-245. 10.1038/nature04559.View ArticlePubMedGoogle Scholar
- Glazko G, Veeramachaneni V, Nei M, Makalowski W: Eighty percent of proteins are different between humans and chimpanzees. Gene. 2005, 346: 215-219. 10.1016/j.gene.2004.11.003.View ArticlePubMedGoogle Scholar
- Yang Z: The power of phylogenetic comparison in revealing protein function. PNAS. 2005, 102 (9): 3179-3180. 10.1073/pnas.0500371102.PubMed CentralView ArticlePubMedGoogle Scholar
- Smith NG, Eyre-Walker A: Human disease genes: patterns and predictions. Gene. 2003, 318: 169-175. 10.1016/S0378-1119(03)00772-8.View ArticlePubMedGoogle Scholar
- Clark AG, Glanowski S, Nielsen R, Thomas PD, Kejariwal A, Todd MA, Tanenbaum DM, Civello D, Lu F, Murphy B, et al: Inferring Nonneutral Evolution from Human-Chimp-Mouse Orthologous Gene Trios. Science. 2003, 302 (5652): 1960-1963. 10.1126/science.1088821.View ArticlePubMedGoogle Scholar
- Huang H, Winter EE, Wang H, Weinstock KG, Xing H, Goodstadt L, Stenson PD, Cooper DN, Smith D, Alba MM, et al: Evolutionary conservation and selection of human disease gene orthologs in the rat and mouse genomes. Genome Biol. 2004, 5 (7): R47-10.1186/gb-2004-5-7-r47.PubMed CentralView ArticlePubMedGoogle Scholar
- Bakewell MA, Shi P, Zhang J: More genes underwent positive selection in chimpanzee evolution than in human evolution. PNAS. 2007, 104 (18): 7489-7494. 10.1073/pnas.0701705104.PubMed CentralView ArticlePubMedGoogle Scholar
- Bustamante CD, Fledel-Alon A, Williamson S, Nielsen R, Hubisz MT, Glanowski S, Tanenbaum DM, White TJ, Sninsky JJ, Hernandez RD, et al: Natural selection on protein-coding genes in the human genome. Nature. 2005, 437 (7062): 1153-1157. 10.1038/nature04240.View ArticlePubMedGoogle Scholar
- Yang Z, Nielsen R: Codon-Substitution Models for Detecting Molecular Adaptation at Individual Sites Along Specific Lineages. Mol Biol Evol. 2002, 19 (6): 908-917.View ArticlePubMedGoogle Scholar
- Zhang J, Nielsen R, Yang Z: Evaluation of an Improved Branch-Site Likelihood Method for Detecting Positive Selection at the Molecular Level. Mol Biol Evol. 2005, 22 (11): 1-8.Google Scholar
- Yang Z, Wong WS, Nielsen R: Bayes empirical bayes inference of amino acid sites under positive selection. Mol Biol Evol. 2005, 22 (4): 1107-1118. 10.1093/molbev/msi097.View ArticlePubMedGoogle Scholar
- Vamathevan J, Holbrook JD, Emes RD: The Mouse Genome as a Rodent Model in Evolutionary Studies. Encyclopedia of Life Sciences. 2007, John Wiley & Sons LGoogle Scholar
- Fraser HB, Hirsh AE, Steinmetz LM, Scharfe C, Feldman MW: Evolutionary rate in the protein interaction network. Science. 2002, 296 (5568): 750-752. 10.1126/science.1068696.View ArticlePubMedGoogle Scholar
- Fraser HB, Wall DP, Hirsh AE: A simple dependence between protein evolution rate and the number of protein-protein interactions. BMC Evol Biol. 2003, 3: 11-10.1186/1471-2148-3-11.PubMed CentralView ArticlePubMedGoogle Scholar
- Fraser HB, Hirsh AE: Evolutionary rate depends on number of protein-protein interactions independently of gene expression level. BMC Evol Biol. 2004, 4: 13-10.1186/1471-2148-4-13.PubMed CentralView ArticlePubMedGoogle Scholar
- Bloom JD, Adami C: Apparent dependence of protein evolutionary rate on number of interactions is linked to biases in protein-protein interactions data sets. BMC Evol Biol. 2003, 3: 21-10.1186/1471-2148-3-21.PubMed CentralView ArticlePubMedGoogle Scholar
- Jordan IK, Wolf YI, Koonin EV: No simple dependence between protein evolution rate and the number of protein-protein interactions: only the most prolific interactors tend to evolve slowly. BMC Evol Biol. 2003, 3: 1-10.1186/1471-2148-3-1.PubMed CentralView ArticlePubMedGoogle Scholar
- Li Y, Wallis M, Zhang YP: Episodic evolution of prolactin receptor gene in mammals: coevolution with its ligand. J Mol Endocrinol. 2005, 35 (3): 411-419. 10.1677/jme.1.01798.View ArticlePubMedGoogle Scholar
- Hao L, Nei M: Rapid expansion of killer cell immunoglobulin-like receptor genes in primates and their coevolution with MHC Class I genes. Gene. 2005, 347 (2): 149-159. 10.1016/j.gene.2004.12.012.View ArticlePubMedGoogle Scholar
- Deeb SS, Jorgensen AL, Battisti L, Iwasaki L, Motulsky AG: Sequence divergence of the red and green visual pigments in great apes and humans. Proc Natl Acad Sci USA. 1994, 91 (15): 7262-7266. 10.1073/pnas.91.15.7262.PubMed CentralView ArticlePubMedGoogle Scholar
- Gibbs RA, Rogers J, Katze MG, Bumgarner R, Weinstock GM, Mardis ER, Remington KA, Strausberg RL, Venter JC, Wilson RK, et al: Evolutionary and biomedical insights from the rhesus macaque genome. Science. 2007, 316 (5822): 222-234. 10.1126/science.1139247.View ArticlePubMedGoogle Scholar
- Arbiza L, Dopazo J, Dopazo H: Positive selection, relaxation, and acceleration in the evolution of the human and chimp genome. PLoS Comput Biol. 2006, 2 (4): e38-10.1371/journal.pcbi.0020038.PubMed CentralView ArticlePubMedGoogle Scholar
- Thomas PD, Kejariwal A, Campbell MJ, Mi H, Diemer K, Guo N, Ladunga I, Ulitsky-Lazareva B, Muruganujan A, Rabkin S, et al: PANTHER: a browsable database of gene products organized by biological function, using curated protein family and subfamily classification. Nucleic Acids Res. 2003, 31 (1): 334-341. 10.1093/nar/gkg115.PubMed CentralView ArticlePubMedGoogle Scholar
- Thomas PD, Campbell MJ, Kejariwal A, Mi H, Karlak B, Daverman R, Diemer K, Muruganujan A, Narechania A: PANTHER: a library of protein families and subfamilies indexed by function. Genome Res. 2003, 13 (9): 2129-2141. 10.1101/gr.772403.PubMed CentralView ArticlePubMedGoogle Scholar
- Online Mendelian Inheritance in Man, OMIM (TM). [http://www.ncbi.nlm.nih.gov/omim]
- Tang K, Thornton KR, Stoneking M: A New Approach for Using Genome Scans to Detect Recent Positive Selection in the Human Genome. PLoS Biol. 2007, 5 (7): e171-10.1371/journal.pbio.0050171.PubMed CentralView ArticlePubMedGoogle Scholar
- Rajagopalan D, Agarwal P: Inferring pathways from gene lists using a literature-derived network of biological relationships. Bioinformatics. 2005, 21 (6): 788-793. 10.1093/bioinformatics/bti069.View ArticlePubMedGoogle Scholar
- Ewan R, Huxley-Jones J, Mould AP, Humphries MJ, Robertson DL, Boot-Handford RP: The integrins of the urochordate Ciona intestinalis provide novel insights into the molecular evolution of the vertebrate integrin family. BMC Evol Biol. 2005, 5 (1): 31-10.1186/1471-2148-5-31.PubMed CentralView ArticlePubMedGoogle Scholar
- Holbrook JD, Sanseau P: Drug discovery and computational evolutionary analysis. Drug Discov Today. 2007, 12 (19–20): 826-832. 10.1016/j.drudis.2007.08.015.View ArticlePubMedGoogle Scholar
- Sachidanandam R, Weissman D, Schmidt SC, Kakol JM, Stein LD, Marth G, Sherry S, Mullikin JC, Mortimore BJ, Willey DL, et al: A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms. Nature. 2001, 409 (6822): 928-933. 10.1038/35057149.View ArticlePubMedGoogle Scholar
- Kaessmann H, Wiebe V, Weiss G, Paabo S: Great ape DNA sequences reveal a reduced diversity and an expansion in humans. Nat Genet. 2001, 27 (2): 155-156. 10.1038/84773.View ArticlePubMedGoogle Scholar
- Varki A: A chimpanzee genome project is a biomedical imperative. Genome Res. 2000, 10 (8): 1065-1070. 10.1101/gr.10.8.1065.View ArticlePubMedGoogle Scholar
- Beniashvili DS: An overview of the world literature on spontaneous tumors in nonhuman primates. J Med Primatol. 1989, 18 (6): 423-437.PubMedGoogle Scholar
- McClure HM: Tumors in nonhuman primates: observations during a six-year period in the Yerkes primate center colony. Am J Phys Anthropol. 1973, 38 (2): 425-429. 10.1002/ajpa.1330380243.View ArticlePubMedGoogle Scholar
- Seibold HR, Wolf RH: Neoplasms and proliferative lesions in 1065 nonhuman primate necropsies. Lab Anim Sci. 1973, 23 (4): 533-539.PubMedGoogle Scholar
- Coggins CR: An updated review of inhalation studies with cigarette smoke in laboratory animals. Int J Toxicol. 2007, 26 (4): 331-338. 10.1080/10915810701490190.View ArticlePubMedGoogle Scholar
- Puente XS, Velasco G, Gutierrez-Fernandez A, Bertranpetit J, King MC, Lopez-Otin C: Comparative analysis of cancer genes in the human and chimpanzee genomes. BMC Genomics. 2006, 7: 15-10.1186/1471-2164-7-15.PubMed CentralView ArticlePubMedGoogle Scholar
- Crespi BJ, Summers K: Positive selection in the evolution of cancer. Biol Rev Camb Philos Soc. 2006, 81 (3): 407-424. 10.1017/S1464793106007056.View ArticlePubMedGoogle Scholar
- Valverde P, Healy E, Sikkink S, Haldane F, Thody AJ, Carothers A, Jackson IJ, Rees JL: The Asp84Glu variant of the melanocortin 1 receptor (MC1R) is associated with melanoma. Hum Mol Genet. 1996, 5 (10): 1663-1666. 10.1093/hmg/5.10.1663.View ArticlePubMedGoogle Scholar
- Lalueza-Fox C, Rompler H, Caramelli D, Staubert C, Catalano G, Hughes D, Rohland N, Pilli E, Longo L, Condemi S, et al: A melanocortin 1 receptor allele suggests varying pigmentation among Neanderthals. Science. 2007, 318 (5855): 1453-1455. 10.1126/science.1147417.View ArticlePubMedGoogle Scholar
- McCallion AS, Chakravarti A: EDNRB/EDN3 and Hirschsprung disease type II. Pigment Cell Res. 2001, 14 (3): 161-169. 10.1034/j.1600-0749.2001.140305.x.View ArticlePubMedGoogle Scholar
- Lahav R: Endothelin receptor B is required for the expansion of melanocyte precursors and malignant melanoma. Int J Dev Biol. 2005, 49 (2–3): 173-180. 10.1387/ijdb.041951rl.View ArticlePubMedGoogle Scholar
- Tascou S, Nayernia K, Uedelhoven J, Bohm D, Jalal R, Ahmed M, Engel W, Burfeind P: Isolation and characterization of differentially expressed genes in invasive and non-invasive immortalized murine male germ cells in vitro. Int J Oncol. 2001, 18 (3): 567-574.PubMedGoogle Scholar
- Katoh M: GIPC gene family (Review). Int J Mol Med. 2002, 9 (6): 585-589.PubMedGoogle Scholar
- Yoon SN, Ku JL, Shin YK, Kim KH, Choi JS, Jang EJ, Park HC, Kim DW, Kim MA, Kim WH, et al: Hereditary nonpolyposis colorectal cancer in endometrial cancer patients. Int J Cancer. 2008, 122 (5): 1077-1081. 10.1002/ijc.22986.View ArticlePubMedGoogle Scholar
- Laganiere J, Deblois G, Lefebvre C, Bataille AR, Robert F, Giguere V: From the Cover: Location analysis of estrogen receptor alpha target promoters reveals that FOXA1 defines a domain of the estrogen response. Proc Natl Acad Sci USA. 2005, 102 (33): 11651-11656. 10.1073/pnas.0505575102.PubMed CentralView ArticlePubMedGoogle Scholar
- Jen JC, Graves TD, Hess EJ, Hanna MG, Griggs RC, Baloh RW: Primary episodic ataxias: diagnosis, pathogenesis and treatment. Brain. 2007, 130 (Pt 10): 2484-2493. 10.1093/brain/awm126.View ArticlePubMedGoogle Scholar
- Jouvenceau A, Eunson LH, Spauschus A, Ramesh V, Zuberi SM, Kullmann DM, Hanna MG: Human epilepsy associated with dysfunction of the brain P/Q-type calcium channel. Lancet. 2001, 358 (9284): 801-807. 10.1016/S0140-6736(01)05971-2.View ArticlePubMedGoogle Scholar
- Loder E: What is the evolutionary advantage of migraine?. Cephalalgia. 2002, 22 (8): 624-632. 10.1046/j.1468-2982.2002.00437.x.View ArticlePubMedGoogle Scholar
- Mahley RW: Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science. 1988, 240 (4852): 622-630. 10.1126/science.3283935.View ArticlePubMedGoogle Scholar
- Hanlon CS, Rubinsztein DC: Arginine residues at codons 112 and 158 in the apolipoprotein E gene correspond to the ancestral state in humans. Atherosclerosis. 1995, 112 (1): 85-90. 10.1016/0021-9150(94)05402-5.View ArticlePubMedGoogle Scholar
- Gearing M, Tigges J, Mori H, Mirra SS: A beta40 is a major form of beta-amyloid in nonhuman primates. Neurobiol Aging. 1996, 17 (6): 903-908. 10.1016/S0197-4580(96)00164-9.View ArticlePubMedGoogle Scholar
- Cole G, Neal JW: The brain in aged elephants. J Neuropathol Exp Neurol. 1990, 49 (2): 190-192. 10.1097/00005072-199003000-00012.View ArticlePubMedGoogle Scholar
- Howlett DR, Richardson JC, Austin A, Parsons AA, Bate ST, Davies DC, Gonzalez MI: Cognitive correlates of Abeta deposition in male and female mice bearing amyloid precursor protein and presenilin-1 mutant transgenes. Brain Res. 2004, 1017 (1–2): 130-136. 10.1016/j.brainres.2004.05.029.View ArticlePubMedGoogle Scholar
- Hacia JG, Fan JB, Ryder O, Jin L, Edgemon K, Ghandour G, Mayer RA, Sun B, Hsie L, Robbins CM, et al: Determination of ancestral alleles for human single-nucleotide polymorphisms using high-density oligonucleotide arrays. Nat Genet. 1999, 22 (2): 164-167. 10.1038/9674.View ArticlePubMedGoogle Scholar
- Strittmatter WJ, Saunders AM, Schmechel D, Pericak-Vance M, Enghild J, Salvesen GS, Roses AD: Apolipoprotein E: high-avidity binding to beta-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease. Proc Natl Acad Sci USA. 1993, 90 (5): 1977-1981. 10.1073/pnas.90.5.1977.PubMed CentralView ArticlePubMedGoogle Scholar
- Emes RD, Pocklington AJ, Anderson CNG, Bayes A, Collins MAO, Vickers CA, Croning MDR, Malik BR, Choudhary JS, Armstrong JD: Evolutionary expansion and anatomical specialization of synapse proteome complexity. Nature Neuroscience.
- Chen Q, Nakajima A, Choi SH, Xiong X, Sisodia SS, Tang YP: Adult neurogenesis is functionally associated with AD-like neurodegeneration. Neurobiol Dis. 2008, 29 (2): 316-326. 10.1016/j.nbd.2007.09.005.PubMed CentralView ArticlePubMedGoogle Scholar
- Finch CE, Morgan TE: Systemic inflammation, infection, ApoE alleles, and Alzheimer disease: a position paper. Curr Alzheimer Res. 2007, 4 (2): 185-189. 10.2174/156720507780362254.View ArticlePubMedGoogle Scholar
- Brune M: Schizophrenia-an evolutionary enigma?. Neurosci Biobehav Rev. 2004, 28 (1): 41-53. 10.1016/j.neubiorev.2003.10.002.View ArticlePubMedGoogle Scholar
- Jungerius BJ, Hoogendoorn ML, Bakker SC, Van't Slot R, Bardoel AF, Ophoff RA, Wijmenga C, Kahn RS, Sinke RJ: An association screen of myelin-related genes implicates the chromosome 22q11 PIK4CA gene in schizophrenia. Mol Psychiatry. 2007Google Scholar
- Stopkova P, Saito T, Papolos DF, Vevera J, Paclt I, Zukov I, Bersson YB, Margolis BA, Strous RD, Lachman HM: Identification of PIK3C3 promoter variant associated with bipolar disorder and schizophrenia. Biol Psychiatry. 2004, 55 (10): 981-988. 10.1016/j.biopsych.2004.01.014.View ArticlePubMedGoogle Scholar
- Saadat M, Pakyari N, Farrashbandi H: Genetic polymorphism in the DNA repair gene XRCC1 and susceptibility to schizophrenia. Psychiatry Res. 2008, 157 (1–3): 241-245. 10.1016/j.psychres.2007.07.014.View ArticlePubMedGoogle Scholar
- Wang X, Baloh RH, Milbrandt J, Garcia KC: Structure of artemin complexed with its receptor GFRalpha3: convergent recognition of glial cell line-derived neurotrophic factors. Structure. 2006, 14 (6): 1083-1092. 10.1016/j.str.2006.05.010.View ArticlePubMedGoogle Scholar
- Kim SJ, Linden DJ: Ubiquitous plasticity and memory storage. Neuron. 2007, 56 (4): 582-592. 10.1016/j.neuron.2007.10.030.View ArticlePubMedGoogle Scholar
- Russo K, Hoch S, Dima C, Varga J, Teodorescu M: Circulating anticentromere CENP-A and CENP-B antibodies in patients with diffuse and limited systemic sclerosis, systemic lupus erythematosus, and rheumatoid arthritis. J Rheumatol. 2000, 27 (1): 142-148.PubMedGoogle Scholar
- Sugimoto K, Migita H, Hagishita Y, Yata H, Himeno M: An antigenic determinant on human centromere protein B (CENP-B) available for production of human-specific anticentromere antibodies in mouse. Cell Struct Funct. 1992, 17 (2): 129-138.View ArticlePubMedGoogle Scholar
- Seeler JS, Muchardt C, Suessle A, Gaynor RB: Transcription factor PRDII-BF1 activates human immunodeficiency virus type 1 gene expression. J Virol. 1994, 68 (2): 1002-1009.PubMed CentralPubMedGoogle Scholar
- Chendrimada TP, Finn KJ, Ji X, Baillat D, Gregory RI, Liebhaber SA, Pasquinelli AE, Shiekhattar R: MicroRNA silencing through RISC recruitment of eIF6. Nature. 2007, 447 (7146): 823-828. 10.1038/nature05841.View ArticlePubMedGoogle Scholar
- Federici T, Boulis NM: Ribonucleic acid interference for neurological disorders: candidate diseases, potential targets, and current approaches. Neurosurgery. 2007, 60 (1): 3-15. 10.1227/01.NEU.0000249214.42461.A5. discussion 15–16View ArticlePubMedGoogle Scholar
- Barnes MR, Deharo S, Grocock RJ, Brown JR, Sanseau P: The micro RNA target paradigm: a fundamental and polymorphic control layer of cellular expression. Expert Opin Biol Ther. 2007, 7 (9): 1387-1399. 10.1517/147125188.8.131.527.View ArticlePubMedGoogle Scholar
- Searls DB: Pharmacophylogenomics: Genes, Evolution and Drug Targets. Nature Reviews Drug Discovery. 2003, 2 (8): 613-10.1038/nrd1152.View ArticlePubMedGoogle Scholar
- Heemskerk J, Tobin AJ, Ravina B: From chemical to drug: neurodegeneration drug screening and the ethics of clinical trials. Nat Neurosci. 2002, 5 (Suppl): 1027-1029. 10.1038/nn931.View ArticlePubMedGoogle Scholar
- Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ: Basic local alignment search tool. J Mol Biol. 1990, 215 (3): 403-410.View ArticlePubMedGoogle Scholar
- Birney E, Clamp M, Durbin R: GeneWise and Genomewise. Genome Res. 2004, 14 (5): 988-995. 10.1101/gr.1865504.PubMed CentralView ArticlePubMedGoogle Scholar
- Consortium ICGS: Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution. Nature. 2004, 432 (7018): 695-716. 10.1038/nature03154.View ArticleGoogle Scholar
- Edgar RC: MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004, 32 (5): 1792-1797. 10.1093/nar/gkh340.PubMed CentralView ArticlePubMedGoogle Scholar
- Murphy WJ, Eizirik E, O'Brien SJ, Madsen O, Scally M, Douady CJ, Teeling E, Ryder OA, Stanhope MJ, de Jong WW, et al: Resolution of the early placental mammal radiation using Bayesian phylogenetics. Science. 2001, 294 (5550): 2348-2351. 10.1126/science.1067179.View ArticlePubMedGoogle Scholar
- Yang Z: PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci. 1997, 13: 555-556.PubMedGoogle Scholar
- Yang Z: Likelihood ratio tests for detecting positive selection and application to primate lysozyme evolution. Mol Biol Evol. 1998, 15 (5): 568-573.View ArticlePubMedGoogle Scholar
- Anisimova M, Yang Z: Multiple Hypothesis Testing to Detect Lineages under Positive Selection that Affects Only a Few Sites. Mol Biol Evol. 2007, 24 (5): 1219-1228. 10.1093/molbev/msm042.View ArticlePubMedGoogle Scholar
- Ingenuity Systems. [http://www.ingenuity.com]
- Jubilant Biosystems. [http://www.jubilantbiosys.com]
- GeneGo. [http://www.genego.com]
- NetPro. [http://www.molecularconnections.com]
- Human Protein Reference Database. [http://www.hprd.org]
- Koike A, Takagi T: PRIME: automatically extracted PRotein Interactions and Molecular Information databasE. Silico Biol. 2005, 5 (1): 9-20.Google Scholar
- Matys V, Fricke E, Geffers R, Gossling E, Haubrock M, Hehl R, Hornischer K, Karas D, Kel AE, Kel-Margoulis OV, et al: TRANSFAC: transcriptional regulation, from patterns to profiles. Nucleic Acids Res. 2003, 31 (1): 374-378. 10.1093/nar/gkg108.PubMed CentralView ArticlePubMedGoogle Scholar
- Kolchanov NA, Ignatieva EV, Ananko EA, Podkolodnaya OA, Stepanenko IL, Merkulova TI, Pozdnyakov MA, Podkolodny NL, Naumochkin AN, Romashchenko AG: Transcription Regulatory Regions Database (TRRD): its status in 2002. Nucleic Acids Res. 2002, 30 (1): 312-317. 10.1093/nar/30.1.312.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.