April 26, 2011 — Many studies have linked brain volume abnormalities to a variety of mental health conditions, including major depressive disorder (MDD), bipolar disorder, and schizophrenia.

But the author of an analysis published online April 4 in Archives of General Psychiatry concludes that the number of statistically significant results in the brain volume literature is "way too large to be true."

Dr. John P. A. Ioannidis

"This pattern suggests strong biases in the literature," writes John P. A. Ioannidis, MD, DSc, of Stanford University's Prevention Research Center in California. "Selective outcome reporting and selective analyses reporting" are 2 possible explanations for the "excess significance bias" uncovered in the literature on brain volume abnormalities.

Reached for comment on the analysis, John D. Port, MD, PhD, associate professor of radiology and assistant professor of psychiatry at Mayo Clinic, Rochester, Minnesota, called Dr. Ioannidis' article and analysis "pretty darn good and not surprising at all." The excess of statistically significant results, Dr. Port said, "is a side effect of statistics and publication bias, in my opinion."

Many Links 'Likely Spurious'

Dr. Ioannidis has published numerous papers on a variety of medical topics that question the methods and interpretations of clinical trials and the ideas they generate. One of his most widely read publication may be the 2005 essay published in PLoS Medicine entitled "Why Most Published Research Findings Are False."

In his latest analysis, he used an "excess significance test" to evaluate whether there are too many reported studies in the brain volume literature that have statistically significant results.

"In a nutshell...there are too many studies done in the field showing too many significant results with brain volume abnormalities [and] many of these significant associations are likely to be spurious," Dr. Ioannidis told Medscape Medical News.

From 8 articles, he evaluated 41 meta-analyses with 461 data sets pertaining to brain volume abnormalities in 7 conditions: major depressive disorder, bipolar disorder, obsessive-compulsive disorder, posttraumatic stress disorder, autism, first-episode schizophrenia, and relatives of patients with schizophrenia.

Of the 41 meta-anlayses, roughly half (n = 21) found "statistically significant" associations, and 142 of the 461 data sets (31%) found a "positive" association.

"Even if the effect sizes observed in the meta-analyses are accurate, the number of positive results (n = 142) is almost double than what would have been expected (n = 78) based on power calculations for the included samples," said Dr. Iaonnidis.

And no condition is spared. On the basis of his research, "bias may be present in meta-analyses of all 7 examined conditions and in most of the examined brain structures," Dr. Ioannidis writes.

"The whole field needs transparent design and reporting of its results and careful reappraisal of putative associations," he said.

'Statistical Noise'

Dr. Port, who was not involved in the analysis, said scientists "need to supply the full data set (positive and negative analyses) and publishers need to be willing to publish the full data set so meta-analyses have the full data to look at."

Publication bias toward the positive findings, he added, "creates, in my opinion, what I like to call statistical noise.

"We know if you set a P value of .05, 1 out of every 20 results is going to be abnormally positive, by definition. So if you are only going to publish positive results, you're going to be publishing some by mistake [noise results], and these show up in these meta-analyses as a bias toward excess significance," Dr. Port explained.

To tackle this issue, "we can start by using more rigorous statistical limits, so instead of using a P value of .05 — use .01 — which means a lot less stuff will be significant, which, unfortunately, means a lot less papers will be published," he added.

The Bonferroni correction can also help. "We hate to do Bonferroni correction because you take a whole bunch of positive results and you get rid of them, but it's a very rigorous test, and if something is really positive, it will survive a Bonferroni correction."

The analysis had no funding. Dr. Ioannidis and Dr. Port have disclosed no relevant financial relationships.

Arch Gen Psychiatry. Published online April 4, 2011.

# posted by roodbont @ 12:38 PM 0 comments

Abstract and Introduction

Abstract

Dementias and related cognitive disorders of the brain are strongly age-associated and prevalence is expected to rise dramatically with a rapidly aging population. As a result, there has been increasing attention on the prevention and treatment of cognitive decline associated with these conditions. A number of approaches have been designed to maintain and strengthen the cognitive capacity of the healthy, as well as the pathologically damaged brain. Evidence suggests that despite advancing age, our brains, and thus our cognitive functions, retain the ability to be maintained and strengthened through the biological process of neuroplasticity. With this opportunity, a new commercial field of 'brain fitness' has been launched to bring to the market training exercises and games that maintain and strengthen cognitive abilities in adulthood. However, the majority of brain fitness methods and products now marketed and sold to consumers have scant scientific evidence to support their effectiveness.

Section 1 of 15

Next: Rationale Behind Brain Exercise to Avoid Cognitive Decline »

# posted by roodbont @ 12:35 PM 0 comments

Abstract and Introduction

Abstract

As the primary target of therapy in the management of dyslipidemia, LDL-C has been a central focus for practicing clinicians for more than a decade. National Cholesterol Education Program guidelines encourage physicians to lower LDL-C levels to outlined therapeutic targets on the basis of ongoing randomized controlled trials demonstrating significant benefit in cardiovascular outcomes among primary and secondary prevention individuals. Relevant epidemiological analysis of cardiovascular outcomes in the USA reports that although statin therapy provides a relative risk reduction of 30%, many coronary heart disease patients at the LDL-C target level are still having major events, of which more than half are recurrent. Although statins – the mainstay of therapy – are able to decrease LDL-C by a range of approximately 30–50% depending on the potency and dose of the statin administered, they remain underused in the clinical setting by practicing physicians. There also remains controversy as to whether more intensive lowering of LDL-C provides additional cardiovascular benefit or not. Intensive lowering of LDL-C as it pertains to the incidence of cardiovascular outcomes (including myocardial infarction, coronary revascularization and ischemic stroke) is assessed in this meta-analysis of 170,000 individuals from 26 large, randomized controlled trials. The implications of the Cholesterol Treatment Trialists' Collaboration for practicing physicians are discussed here.

# posted by roodbont @ 9:53 AM 0 comments

SUMMARY AND COMMENT

Statin Use and Risk for Death After Pneumonia

April 28, 2011 | Paul S. Mueller, MD, MPH, FACP

Statin users were significantly less likely to die in the 6 months after pneumonia diagnoses.

Reviewing: Douglas I et al. BMJ 2011 Apr 6; 342:d1642

# posted by roodbont @ 7:13 AM 0 comments

Abstract

Alzheimer's disease (AD), the most common cause of dementia, is a progressive neurodegenerative disorder affecting millions of people worldwide. AD has a multifactorial origin, resulting from an interaction between genetic susceptibility and environmental risk factors. Genetic, epidemiological, experimental and clinical data strongly suggest that the metabolism of cholesterol has an important role in AD pathogenesis. Several studies have demonstrated that high concentrations of serum cholesterol increase the risk of AD. Statins, drugs that reduce cholesterol levels, have been investigated as a possible treatment for AD. However, the literature is not exempt of contradictory results. In this article, we review a recent article by Reitz et al. demonstrating that higher levels of high-density lipoprotein cholesterol, total cholesterol and non-high-density lipoprotein cholesterol are associated with lower risk of AD. In addition, we discuss the current state of knowledge regarding the relationship between plasma cholesterol and AD, stressing the need for understanding the molecular mechanisms behind this association.

Introduction

Alzheimer's disease (AD) is a progressive neurodegenerative disorder that constitutes the main cause of dementia in the elderly. Two well-established risk factors for AD are age and the presence of the ε4 isoform of apolipoprotein E (apoE). In the past decade, several risk factors have been associated with AD, including hypertension, diabetes mellitus, obesity, metabolic syndrome and hypercholesterolemia.^[1] In addition, cholesterol metabolism has been previously implicated in AD pathogenesis.^[2–4] However, epidemiological studies have reported contradictory results concerning the role of cholesterol as a risk factor for AD. Globally, a positive association with AD was found when cholesterol levels were measured in middle-aged individuals,^[5] with the exception of the Honolulu Asia Study, where a cluster of cardiovascular factors increased the risk of vascular dementia (VaD) but not of AD.^[6]

Epidemiological studies demonstrated that apoE4 carriers have a higher risk for both AD and an early onset of the disease.^[7] Genetic studies have identified several cholesterol-related genes that are associated with AD.^[8] It has been suggested that both genetic and environmental risk factors could act in synergy to confer the risk for AD.^[9]

Based on evidence supporting hypercholesterolemia as a risk factor for AD, cholesterol synthesis inhibitors, such as statins, were proposed as valuable tools for therapeutic intervention. Despite the fact that the majority of retrospective studies with statins demonstrated a positive effect, others were not able to determine an association between the use of these drugs, cognitive decline and dementia or AD risk. It has been proposed that the different capacity of statins to pass the BBB, their different pleotropic effects, or the necessity to start treatment in very early phases of the disease may have caused these controversial results.^[10–12]

Vascular dysfunction in the brain has long been recognized as a contributing factor to the development of VaD. More recently, accumulated evidence suggests that the development of AD may also be strongly related to underlying vascular problems.^[13] Furthermore, it seems obvious to infer that a disturbance in plasma cholesterol levels or metabolism could participate in vascular dysfunctions, and therefore in the development of VaD.

The molecular mechanisms linking cholesterol with AD pathology are still unknown. Amyloidβ (Aβ) accumulation in the brain is thought to play a key role in the neurodegenerative processes and results from both brain overproduction and aberrant clearance of these peptides across the BBB.^[14] In vitro studies have also demonstrated that intracellular cholesterol levels can modulate the processing of APP to Aβ and that the cleavage of APP by β- and γ-secretases occurs in cholesterol-rich microdomains within the neuronal membranes known as 'lipid rafts'. On the other hand, cholesterol metabolites were also demonstrated to modulate APP processing^[15] and Aβ production.

Since the brain is separated from the plasma cholesterol by the BBB, it is intriguing how high serum cholesterol levels influence the prevalence of AD. Oxysterols are capable of passing through the BBB, and high levels of 27-hydroxycholesterol passing from the blood into the brain has been proposed as a possible mechanism.^[16] Recent data demonstrated that high cholesterol and 27-hydroxycholesterol levels affect memory consolidation in experimental studies^[17,18] and are intimately linked to the renin–angiotensin system overactivation observed in the brains of AD patients.^[19]

Although midlife studies consistently associate high plasma cholesterol levels with dementia, this association is less clear in older people. Reitz et al. reported that high levels of high-density lipoprotein cholesterol (HDL-C) in elderly individuals are associated with a decreased risk of late-onset AD.^[20] This result, which may seem controversial, has also been reported in other previous studies.^[21]

In this article, we discuss possible explanations for these apparently controversial findings, and emphasize the need to understand how blood lipid levels participate in the pathogenetic mechanisms leading to different dementia-causing disorders.

Section 1 of 4

Next: Methods & Results »

# posted by roodbont @ 3:24 PM 0 comments

Abstract

As the primary target of therapy in the management of dyslipidemia, LDL-C has been a central focus for practicing clinicians for more than a decade. National Cholesterol Education Program guidelines encourage physicians to lower LDL-C levels to outlined therapeutic targets on the basis of ongoing randomized controlled trials demonstrating significant benefit in cardiovascular outcomes among primary and secondary prevention individuals. Relevant epidemiological analysis of cardiovascular outcomes in the USA reports that although statin therapy provides a relative risk reduction of 30%, many coronary heart disease patients at the LDL-C target level are still having major events, of which more than half are recurrent. Although statins – the mainstay of therapy – are able to decrease LDL-C by a range of approximately 30–50% depending on the potency and dose of the statin administered, they remain underused in the clinical setting by practicing physicians. There also remains controversy as to whether more intensive lowering of LDL-C provides additional cardiovascular benefit or not. Intensive lowering of LDL-C as it pertains to the incidence of cardiovascular outcomes (including myocardial infarction, coronary revascularization and ischemic stroke) is assessed in this meta-analysis of 170,000 individuals from 26 large, randomized controlled trials. The implications of the Cholesterol Treatment Trialists' Collaboration for practicing physicians are discussed here.

# posted by roodbont @ 3:21 PM 0 comments

Cryoballoon ablation more effective for paroxysmal than persistent AF

# posted by roodbont @ 3:23 PM 0 comments

Lengthened WHI Follow-Up: Postmenopausal Estrogen Therapy

Women's Health Initiative follow-up of postintervention phase showed estrogen had no substantial adverse effects on most health outcomes; reduction in relative risk for breast cancer persisted.

In the Women's Health Initiative (WHI) Estrogen-Alone Trial, 11,000 postmenopausal women with hysterectomies (age range at baseline, 50–79) received conjugated equine estrogen or placebo for a median of 5.9 years; the trial was stopped at mean follow-up of 7.1 years when initial findings showed excess risk for stroke in the estrogen group. Subsequent analysis showed that, in younger WHI participants (age range at baseline, 50–59), estrogen use was associated with lower risk for coronary heart disease (CHD) and overall mortality during the intervention period. Now, investigators have assessed postintervention health outcomes in 7645 WHI participants.

Overall, at a mean 10.7 years after baseline, estrogen use was not associated with excess risk for stroke or other adverse outcomes (CHD, deep venous thrombosis, hip fracture, colorectal cancer, and mortality during follow-up). As had been noted in previous WHI analyses of the intervention phase, risk for invasive breast cancer was lower in the estrogen group than in the placebo group (hazard ratio, 0.77; 95% confidence interval, 0.62–0.95); moreover, among younger women, estrogen use was associated with lower risk for CHD (HR, 0.59; 95% CI, 0.38–0.90) and marginally lower risk for mortality during follow-up (HR, 0.73; 95% CI, 0.53–1.00).

Comment: Although these new findings from the WHI are intriguing, they do not imply that estrogen-only therapy should be recommended at this time for cardioprotection or breast cancer chemoprophylaxis. What these results do provide is reassurance to young menopausal women who are posthysterectomy and who present with bothersome vasomotor symptoms that use of estrogen therapy for as long as 6 years is safe.

— Andrew M. Kaunitz, MD

Published in Journal Watch Women's Health April 5, 2011

# posted by roodbont @ 12:49 AM 0 comments

SUMMARY AND COMMENT

Long-Term Statin Therapy Does Not Lower Risk for Atrial Fibrillation

April 7, 2011 | Paul S. Mueller, MD, MPH, FACP

Several factors might explain discrepant results between short-term and long-term trials.

Reviewing: Rahimi K et al. BMJ 2011 Mar 16; 342:d1250

# posted by roodbont @ 12:37 AM 0 comments

Is the Radial Approach for Everyone? There are a few prerequisites for patients to be a candidate for the transradial approach. The first is confirmation of a dual, or "protected", blood supply to the hand. The radial artery loops around the hand and joins the ulnar artery. Both arteries supply blood to the hand and fingers. It is precisely this dual blood supply that makes the radial technique safe. Should the radial artery close up (a complication seen in a small percentage of cases) the clinical result tends to be benign, because the ulnar artery continues to function. The first step a cardiologist takes in deciding on the radial approach is an Allen test to assess that both radial and ulnar arteries are functioning normally -- a simple test that can be done by compressing the arteries by hand at bedside or in the doctor's office. If they are not normal, then the femoral approach is preferred. Some other contraindications exist, such as the need to use larger devices during the angioplasty, pre-existing bypass grafts in certain areas or tortuous vessels that may prevent the catheter from navigating to the coronaries from the arm. About 30-40% of patients are not candidates for radial access.

While the complication rate with the radial approach is extremely low, there is always some risk with any medical procedure. It is important for patients to discuss the risks and benefits of the femoral vs. radial approaches, as these can vary for each individual.

# posted by roodbont @ 12:44 AM 0 comments

Another Contender in the Race to Unseat Warfarin

Mark S. Link, MD

Authors and Disclosures

Posted: 03/25/2011; Journal Watch © 2011 Massachusetts Medical Society

Abstract and Introduction

Abstract

Compared with aspirin, apixaban reduced the risk for embolic events in patients with atrial fibrillation.

Introduction

A fierce competition is under way to develop a replacement for warfarin in the treatment of atrial fibrillation (AF). Dabigatran, a direct thrombin inhibitor, is approved for use in the U.S., and rivaroxaban, a factor Xa inhibitor, was noninferior to warfarin in the preliminary results of the ROCKET AF trial. Now, apixaban, another factor Xa inhibitor, has been compared with aspirin in an industry-sponsored trial.

The AVERROES investigators randomized 5599 patients with AF and at least one additional risk factor for stroke to apixaban (5 mg twice daily) or aspirin (81–324 mg daily). All patients were considered unsuitable for warfarin treatment, 40% because of prior problems with the drug.

The study was terminated early because of demonstrated superiority of apixaban; mean follow-up was 1.1 years. The rate of the primary outcome — stroke or systemic embolism — was 1.6% per year in the apixaban group versus 3.7% per year in the aspirin group (hazard ratio, 0.45; P<0.001). Major bleeding rates were similar in the two groups (apixaban, 1.4% per year; aspirin, 1.2% per year). The rate of the composite of stroke, systemic embolism, myocardial infarction, death from vascular causes, and major bleeding was 5.3% per year in the apixaban group versus 7.2% per year in the aspirin group (HR, 0.74; P=0.003).

# posted by roodbont @ 12:43 PM 0 comments

From Future Neurology

Dabigatran for Stroke Prevention in Patients with Atrial Fibrillation and Previous Stroke or Transient Ischemic Attack

Abstract and Introduction

Abstract

This study aimed to assess the efficacy and safety of daibgatran in two doses (110 and 150 mg) compared with warfarin in a prespecified subgroup analysis of patients with previous stroke or transient ischemic attack in the Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) trial. There were nonsignificant risk reductions for the primary outcome (stroke and systemic embolism) for both doses of dabigatran compared with warfarin in this subgroup of patients. However, the 110-mg dose of dabigatran provided significantly greater reductions of mortality and higher net clinical benefit compared with warfarin. This was not seen in the 150-mg dose. The bleeding complication rates of this subgroup were consistent with the main RE-LY trial. In the warfarin group, patients with previous history of stroke or transient ischemic attack developed more intracranial bleeding than patients without this history, but this was not the case in dabigatran treatment groups.

# posted by roodbont @ 12:40 PM 0 comments

Abstract

Atrial fibrillation (AF) is associated with significant morbidity and mortality. It is also a progressive disease secondary to continuous structural remodelling of the atria due to AF itself, to changes associated with ageing, and to deterioration of underlying heart disease. Current management aims at preventing the recurrence of AF and its consequences (secondary prevention) and includes risk assessment and prevention of stroke, ventricular rate control, and rhythm control therapies including antiarrhythmic drugs and catheter or surgical ablation. The concept of primary prevention of AF with interventions targeting the development of substrate and modifying risk factors for AF has emerged as a result of recent experiments that suggested novel targets for mechanism-based therapies. Upstream therapy refers to the use of non-antiarrhythmic drugs that modify the atrial substrate- or target-specific mechanisms of AF to prevent the occurrence or recurrence of the arrhythmia. Such agents include angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), statins, n-3 (ω-3) polyunsaturated fatty acids, and possibly corticosteroids. Animal experiments have compellingly demonstrated the protective effect of these agents against electrical and structural atrial remodelling in association with AF. The key targets of upstream therapy are structural changes in the atria, such as fibrosis, hypertrophy, inflammation, and oxidative stress, but direct and indirect effects on atrial ion channels, gap junctions, and calcium handling are also applied. Although there have been no formal randomized controlled studies (RCTs) in the primary prevention setting, retrospective analyses and reports from the studies in which AF was a pre-specified secondary endpoint have shown a sustained reduction in new-onset AF with ACEIs and ARBs in patients with significant underlying heart disease (e.g. left ventricular dysfunction and hypertrophy), and in the incidence of AF after cardiac surgery in patients treated with statins. In the secondary prevention setting, the results with upstream therapies are significantly less encouraging. Although the results of hypothesis-generating small clinical studies or retrospective analyses in selected patient categories have been positive, larger prospective RCTs have yielded controversial, mostly negative, results. Notably, the controversy exists on whether upstream therapy may impact mortality and major non-fatal cardiovascular events in patients with AF. This has been addressed in retrospective analyses and large prospective RCTs, but the results remain inconclusive pending further reports. This review provides a contemporary evidence-based insight into the role of upstream therapies in primary (Part I) and secondary (Part II) prevention of AF.

# posted by roodbont @ 12:37 PM 0 comments

BMC Biology 2011, 9:20doi:10.1186/1741-7007-9-20

The electronic version of this article is the complete one and can be found online at: http://www.biomedcentral.com/1741-7007/9/20

© 2011 Liang and Nielsen; licensee 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.

Is it true that modern humans have Neanderthals and other archaic species in their direct ancestry?

According to two recently published papers by Green et al. and Reich et al., the answer to this question is yes. Human genomes are in part composed of DNA from other archaic hominin species that traditionally have not been counted among our ancestors, although the proportion of archaic DNA in the genome depends on your ethnicity. On the basis of analyses of ancient DNA, Green et al. report that, on average, 1 to 5% of the genomes of non-African individuals are descended from a Neanderthal, and Reich et al. report that 4 to 6% of the genomes of Melanesians are derived from a newly discovered archaic hominin population dubbed the Denisovans. Denisovans and Neanderthals are the only archaic species investigated so far, but future investigations may reveal contributions of DNA from other species, perhaps even from species that have never been characterized well morphologically.

What is an archaic hominin, exactly?

Hominins are humans and their closely related extinct ancestors. Denisovans and Neanderthals were hominins that last lived approximately 30,000 years ago. Neanderthal fossils were first found in 1856, in the Neander Valley, which lends its name to the species. Since then, specimens have been found in a wide geographical range, including the Middle East, Central Asia, and Western and Central Europe. To date, the only discovered Denisovan remains are the finger bone and two teeth discovered in Denisova Cave in Siberia. On the basis of genetic analysis of the finger bone, Reich et al. conclude that Denisovans represent a deeply diverged population distinct from other Neanderthals. Whether Neanderthals and Denisovans comprise separate species is probably mainly an issue of semantics and, in any case, cannot be answered without additional Denisovan samples.

How does this fit with current theories of human origins?

The question of human origins has intrigued scientists ever since Darwin first proposed the theory of evolution. Historically, most of the debate has focused on two competing hypotheses: the out of Africa (OOA) theory (Figure 1a) and the multi-regional theory (Figure 1b). The OOA hypothesis posits that anatomically modern humans first evolved in Africa 200,000 to 150,000 years ago and then migrated out of Africa 100,000 to 60,000 years ago, displacing other archaic hominins, and giving rise to all current human populations. The multi-regional theory suggests that archaic hominins spread out of Africa much earlier, and that humans then evolved from this Eurasia-wide population, with some degree of interbreeding, and thus gene flow, among individuals from different populations being responsible for the degree of genetic differentiation between populations we currently observe. Mitochondrial (mt) DNA data first reported in 1987 and subsequent analyses of autosomal DNA seemed to support the OOA hypothesis.

However, even before the publication of the Neanderthal genome, analyses of modern human DNA from different geographic sources by Jeffrey Wall and others had suggested that, contrary to the earlier consensus, anatomically modern humans evolved in Africa recently, but admixed with endemic archaic hominids - Neanderthals, Denisovans, or even Homo erectus - as they spread throughout the world (Figure 1c), and that ancestral admixture may be much more common than previously thought.

Wall et al. based their analysis on the pattern of haplotype lengths. After controlling for other confounding factors, such as demographic history and recombination rate variation, they concluded that the observed lengths of these regions could only be accounted for by archaic admixture on the order of 5%. The evidence of admixture from Neanderthal and Denisovan nuclear DNA lends credence to these claims.

Isn't it extremely difficult to get authentic undamaged DNA from individuals dead for over 30,000 years?

Yes. It was necessary to locate samples that had been buried in cool and dry conditions, under which DNA is degraded relatively slowly. Even so, for the Neanderthal samples, most DNA fragments were very short, and approximately 95 to 99% of the DNA in the samples belonged to bacteria. To reduce the amount of sequencing needed, the relative proportion of hominin DNA was increased by treating the DNA extract with a concoction of restriction enzymes that were chosen to cut bacterial DNA preferentially. This increased the relative proportion of Neanderthal DNA to over 10%. This enriched extract was analyzed using new-generation sequencing machines, which produced a draft sequence with approximately 1.3× coverage - that is, on average, each base pair in the genome was sequenced 1.3 times. Because of the random nature of next-generation sequencing, this means that certain parts of the genome will not have been sequenced at all, while other parts will have been sequenced many more times.

The genetic material from the Denisovan individual was extracted from a finger bone. Because of the cooler climate in Siberia, there was less environmental degradation of the DNA. However, the small volume of the finger bone yielded only enough DNA to sequence the Denisovan genome to 1.9× coverage.

We must have a lot of DNA in common with archaic hominins because of our shared ancestry - How can we infer interbreeding?

If we have two human populations, one of which has undergone more archaic admixture than the other, then we expect the more admixed human population to be more genetically similar to the archaic hominin than the other human population. This intuition is formalized by the ABBA-BABA test. In this statistical test, DNA representing the same sites in a chimpanzee sequence, an archaic hominin sequence, and sequences from a pair of modern human populations, such as Han and Yoruban or Japanese and French, designated H1 and H2, are compared. Only sites with two alleles, A and B, are considered. The chimpanzee is assumed to carry A, the ancestral allele. Two numbers are then computed, n_ABBA, the number of sites where the chimpanzee and one of the pair of modern humans (H2) have allele A and the archaic hominin and the other modern human (H1) have allele B (ABBA) and n_BABA, the number of sites where chimpanzee and H1 have allele A and the archaic hominin and H2 have allele B (BABA). Finally, n_ABBA and n_BABA are added up over all pairings of H1/H2 samples from the two human populations being analyzed. If there had been no archaic admixture, then the difference of these sums is expected to be 0. If the difference is significantly different from 0, then the null hypothesis of no admixture is rejected. Using population genetic models, the admixture fraction can also be estimated from the magnitude of this difference. The ABBA-BABA test can then be used for each pair of human populations to determine the differences in admixture rates between them.

The results of the ABBA-BABA test showed that human non-African populations are more closely related to Neanderthals than African populations. When applied to the Denisovan genome, Reich et al. found that only Melanesians showed evidence of admixture.

This seems quite a subtle test - Might these results be explained by human contamination?

Probably not. Contamination is a serious problem in any sequencing project. A recent paper by Longo et al. reports significant human contamination in non-primate genome databases, and previous analyses of Neanderthal genetic material have also been plagued by human DNA contamination.

In the light of this earlier experience, researchers took several precautions to guard against contamination. The initial sample preparation and DNA extraction were done in a clean room, using several procedures to reduce the chances of modern human DNA contamination. As an additional step in the sample preparation, special primers were ligated onto both ends of each fragment, identifying the fragments. During the sequencing, only reads with this clean room tag were used to assemble the draft genome, minimizing the effect of post-clean room contamination.

The efficacy of these methods was validated using three different procedures: by looking at mtDNA; by looking for Y chromosome sequences; and by using statistical analyses of autosomes. mtDNA is much easier to sequence because it occurs in much higher concentration than nuclear DNA. As a result, the Neanderthal mtDNA sequence can be very accurately determined, and several fixed differences between humans and Neanderthals have been identified. These differences can be used to estimate the ratio of human mtDNA to Neanderthal mtDNA in the sample. Likewise, because all the samples were female, the amount of Y chromosomal DNA can be used to estimate the level of contamination from human males. Finally, researchers used human heterozygosity and allele frequency data to directly estimate contamination in the autosomal DNA. All three methods estimated the human contamination to be 1% or less.

This is consistent with the results of a blind test in which Green et al. examined present-day human genetic variation without knowledge of the Neanderthal sequence, and were able to locate regions of the human genome that appeared admixed. Comparison of their predictions with the Neanderthal data showed that these candidate regions matched the Neanderthal sequence at a higher frequency than could be explained by any level of contamination.

What about DNA damage?

The main problem in dealing with ancient DNA is the dearth of genetic material. The Neanderthal and Denisovan genomes could not be sequenced to a higher coverage not because of a lack of money or time, but because of a lack of DNA extract; the three bones from Vindija Cave and the one from Denisova Cave have been completely hollowed out to produce the genomes reported.

Ancient DNA sequencing typically shows a much high error rate than observed in modern DNA. Errors in the reported genome can be caused by degradation of the DNA from the environment or by sequencing error. In ancient DNA samples, deamination of cytosine residues causes C to have the chemical properties of T, and G to have the chemical properties of A. As a result, the Neanderthal draft genome shows an abnormally large number of C-> T and G-> A substitutions, the vast majority of which are errors. In sequencing the Denisovan samples, this deamination was chemically reversed, allowing the C and G residues to be sequenced correctly. This, together with the drier and cooler climate at Denisova Cave, resulted in DNA samples that were about ten times less damaged.

Sequencing error can also be a problem, as the error rate of new-generation sequencing is only slightly lower than the divergence between humans and Neanderthals. However, this problem will hopefully disappear as new-generation sequencing technology becomes more accurate and the discovery of new samples allows for deeper coverage.

That sounds serious - How confident can we be of any interpretation if the sequencing error rate and the divergence are that close?

The statistical analysis of the Neanderthal and Denisovan genomes was designed with the limitations of the data in mind. A paper by Durand et al. argues that the ABBA-BABBA test for admixture is not sensitive to confounding factors, such as human or Neanderthal demographic history, sequencing error or damage to the DNA, as long as the H1 and H2 samples were processed in the same way. However, one source of concern is the possibility of a shared error structure caused by DNA sequencing methods. Current sequencing technology is highly temperamental, and the frequency and type of sequencing errors in the final data depend on many factors, such as sample preparation, the type of sequencing machine, contamination from local conditions and reagents, and sequencing coverage. If the error structures of the archaic DNA and one of the modern human DNA samples are similar to each other for one of many reasons, the ABBA-BABA test could report admixture when it did not in fact occur. Even a very small proportion of shared errors could cause a strong effect on the ABBA-BABA statistic. For example, small effects that we typically tend to ignore, such as shared contamination of reagents between the samples, could cause artifactual evidence of admixture. Green et al. and Reich et al. made great efforts to control for these effects, and appear to have succeeded. However, the issues of errors in next-generation sequencing data, particularly for ancient DNA, and their consequences for current and future inference of low levels of admixture remain a critical issue that is likely to be the focus of much future research.

Assuming that we can be confident of the conclusions of these studies, how much of our genomes comes from other hominins?

These two papers only investigated the possibility of admixture from Neanderthals and Denisovans into humans. It is possible that other archaic hominins, perhaps as yet undiscovered, also contributed to the human genome. In fact, Plagnol and Wall report that there is evidence for significant admixture into African populations as well, although no candidate species has been proposed.

On the basis of the data and analyses presented by Green et al. and Reich et al., it appears that a simple out of Africa hypothesis with no admixture does not give the full picture of human origins. As sequencing technology improves and additional archaeological discoveries are made, we should be able to gain a more detailed understanding of what now seems to be the mosaic ancestry of the human genome.

Where can I find out more?

1. Green RE, Krause J, Briggs AW, Maricic T, Stenzel U, Kircher M, Patterson N, Li H, Zhai W, Fritz MH, Hansen NF, Durand EY, Malaspinas AS, Jensen JD, Marques-Bonet T, Alkan C, Prüfer K, Meyer M, Burbano HA, Good JM, Schultz R, Aximu-Petri A, Butthof A, Höber B, Höffner B, Siegemund M, Weihmann A, Nusbaum C, Lander ES, Russ C, et al.: A draft sequence of the Neandertal genome. Science 2010, 328:710-722.

2. Reich D, Green RE, Kircher M, Krause J, Patterson N, Durand EY, Viola B, Briggs AW, Stenzel U, Johnson PL, Maricic T, Good JM, Marques-Bonet T, Alkan C, Fu Q, Mallick S, Li H, Meyer M, Eichler EE, Stoneking M, Richards M, Talamo S, Shunkov MV, Derevianko AP, Hublin JJ, Kelso J, Slatkin M, Pääbo S: Genetic history of an archaic hominin group from Denisova Cave in Siberia. Nature 2010, 468:1053-1060.

3. Durand E, Patterson N, Reich D, Slatkin M: Testing for ancient admixture between closely related species. Genetics 2011, in press.

4. Wall J, Hammer M: Archaic admixture in the human genome. Curr Opinin Genet Dev 2006, 16:606-610.

5. Plagnol V, Wall J: Possible ancestral structure in human populations. PLoS Genet 2006, 2:972-979.

# posted by roodbont @ 12:41 PM 0 comments