We are searching data for your request:
Upon completion, a link will appear to access the found materials.
How is it possible for snails in Ireland to be genetically the same as snails in France? Well probably they didn’t travel thousands of miles including sea on their own. Scientists believe that this happened due to a human migration going back 8000 years.
The similarity of the genetic material, in addition to the fact that those snails cannot be found in Britain, suggests that their location is related to human migration. A research paper published on PLOS ONE by Adele J. Grindon and Angus Davinson explains in detail why this is possible.
Pyrenean glass snails are not the only species found in Ireland in an unexplained way. It also includes a strawberry tree and the Kerry slug. The samples that were collected were from France, Spain and Italy as well as the west coast of Ireland. DNA was extracted from the samples and compared to each other showing the similarities in the genetic material of all samples.
And while Britain is the main origin of the Irish fauna, in this case this doesn’t apply. The explanation that someone would give is that snails migrated from the Pyrenees mountains to France and then to Britain and then to Italy. However the problem with this theory is that traces of the specific genetic material cannot be found in Britain.
Therefore, Davinson suggests that the evidence shows the possibility that Ireland was colonized earlier than Britain, at at least 7000 years BC, bringing with them snails from mainland Europe. We know that snails have been used as a source of food in prehistoric times (back to 9,000 BC) and even farmed for that reason, which makes the suggestion plausible.
Tracking Human Migrations through … snails - History
Migration is one of the great forces of history. When people move in large numbers they sometimes rearrange not only their own lives but also places they leave and the places they settle. Americans have always been a moving people, coming from other places, moving to new places. Not only has the nation long attracted people from other nations, it also claims high rates of internal migration. This project explores a number of consequential migrations--Great Migrations--that helped reshape culture, politics, or economic structures. It has five units, each with detailed information and interactive maps, charts, and data: (1) the migration of African Americans out of the South 1900-2000 (2) the enormously consequential migrations of Latinx Americans, both from Latin America and inside the US (1850-2017) (3) the diaspora of whites from the South to northern and western states (4) the Dust Bowl migration to California from Oklahoma and neighboring states in the 1930s. (5) In addition, we provide migration histories for all fifty states showing decade-by-decade from 1850-2017 where residents have come from.
Most of these materials are from published and unpublished work by James Gregory, Professor of History, University of Washington.
Upwards of 7 million African Americans left the South during the 20th century, settling mostly in the big cities of the North and West. In doing so they transformed more than their own lives. This Great Migration transformed cities and set the foundations for reconstructions of race, politics, and even the regional balances of the nation. This section includes six interactive maps and charts and several interpretative essays.
More than 20 million whites left the South during the 20th century, vastly outnumbering the 7-8 African Americans who left. They were joined by nearly 1 million Latinx, mostly Tejanos, who moved west to California and north into the Midwest. This section shows migration patterns and explores the impacts of the southern diaspora. This section includes six interactive maps and tables as well interpretative essays.
The relocation to California of close to 400,000 Oklahomans, Texans, Arkansans, and Missourians during the Great Depression was the most publicized mass migration of that decade. Many faced unexpected difficulties, especially those who headed for California's Central Valley. Their plight caught the attention of journalists, photographers, and became the subject of one of the most celebrated American novels of the century, John Steinbeck's The Grapes of Wrath. Here are three interactive maps as well as detailed accounts and primary sources.
Americans have always been a moving people, coming from other places, moving to new places. Not only has the nation long attracted people from other nations, it also claims high rates of internal migration. Here are interactive graphics and maps that allow us to track the changing population decade-by-decade since 1850. Select a state and see where people were born, both other countries and other states. We have more detail about key states, including racial breakdowns, see the separate pages for: Alabama, Arizona, California, Colorado, Florida, Illinois, New York, Texas, Virginia, Washington state,.
California Migration History 1850-2017
California's history is keyed to migration. The most populous state in the union became so because so many people from other states and other lands have moved there. It was not until 2010 that the number of native-born Californians surpassed the number who had migrated from somewhere else. And still today most adults are from another state or another country. Migration predated the period of US control notably when Spain sent soldiers and missionaries into the area they named California. It accelerated after the United States seized the Mexican province and immediately profited from the 1848 discovery of gold in the Sierra foothills. [more]
Florida Migration History 1850-2017
More than any other southern state, Florida has consistantly attracted newcomers. Today only about one third of the population claims a Florida birthplace two thirds are from somewhere else, many from Cuba and Latin America but really from every state and most nations. Migration to Florida is an old story, In the late 1800s, Florida grew rapidly as whites from neighborhing states moved south looking for land to farm. [more]
Illinois Migration History State 1880-2017
Illinois, like other midwestern states, experienced rapid population growth through migration in the 19th century and much slower growth since then. Recording a population of 851,000 in 1850, the state doubled its numbers by 1860, doubled again by 1880, and again before 1910. It has taken a century to double the 1910 total. Migration in the 19th century drew heavily on states to the east of Illinois, especially New York, Ohio, and Pennsylvania. European immigrants from Gemany and Ireland came in huge numbers in the same period. The 1880 census showed that roughly half the population were from out of state with Germany, New York, Ohio, and Ireland the leading contributors. [more]
New York Migration History 1850-2017
New York has always been the gateway state, the state that absorbs the greatest diversity of newcomers from abroad. In 1850 when for the first time the US census recorded birthplaces, the leading birthplaces for residents not born in New York were in order: Ireland, Germany, England, Connecticut, Massachusetts, Vermont, New Jersey, and Canada, with a dozen other countries further down the list. [more]
Texas Migration History 1850-2017
Texas has been a migration magnet throughout its history, which helps explain the record of growth that now makes it the second most populous state following California. Migration, most of it illegal, from Tennessee, Alabama, and Mississippi fueled the rebellion that wrestled the province from Mexico in 1836. In the decades that followed an enormous number of southerners moved west to expand the cotton belt, many of them enslaved. [more]
Washington Migration History 1850-2017
Washington remains today a state where most residents came from somewhere else, another state or another country. Occupied by Americans since the 1840s, it's population grew slowly until statehood in 1889, then surged after the discovery of gold in Alaska and the Yukon in 1896. Farming, timber, and fishing attracted newcomers from the upper Midwest and from northern Europe, along with smaller numbers of Chinese and Japanese. [more]
Arizona Migration History 1850-2017
Arizona remains today a state where most residents came from somewhere else, either another state or another country. It is also a state where substantial tracks of land remain the property of Native nations--Navaho, Yuma, Pima, Apache, Pueblo, Papago, and Zuni. The United States seized the area in the war against Mexico in 1846, but few Americans found reason to settle there until silver and cooper deposits brought miners starting in the late 1870s. An 1870 population of less than 10,000 grew to 122,000 by 1900. Whites comprised less than half of the turn-of-the-century population. One third were ethnic Mexicans, born either in the Southwest or Mexico. Native peoples accounted for more than 20 percent of persons enumerated in the 1900 census. [more]
Colorado Migration History 1850-2017
Colorado remains today a state where most residents came from somewhere else, either another state or another country. Cheyenne, Shoshone, Arapahoe, Ute, and other native nations commanded the area until 1848 when the United States seized it in its war against Mexico. The discovery of gold near Pike's Peak in 1857 set up the first significant influx of newcomers, mostly whites from Midwestern and Northeastern states. [more]
Virginia Migration History 1850-2017
Until World War II, Virginia shared the demographic and migration patterns common to Southeastern states. The state attracted few newcomers aside from those moving short distances across the borders of continguous states. Natural increase among African American and white Virginians drove population gains while outmigration exceeded all avenues of inmigration. World War II brought military installations and defense industries that began to reshape the economy,[more]
Alabama Migration History 1850-2017
Alabama was a population replacement zone in the early 19th century as Choctaw, Creek, Chickisaw, and Cherokee people were driven west and their lands were sold off to White planters from Georgia, South and North Carolinia, Tennessee, and Virgina bent on expanding the cotton kingdom. Enslaved people of African descent did the work and comprised 45 percent of Alabama's population on the eve of the Civil War. Freedom turned Alabama in a different direction. For the next century, until 1960s, few people moved into the state and population grew slowly, largely dependent upon "natural increase," an odd euphemism for the work of mothers. [more]
How to read a Y
Most chromosomes in the cell are present as two copies, which allows them to swap genetic material. Over time, this swapping will mix up the mutations that occur on the chromosome, making their history difficult to untangle.
The Y chromosome is different in that males only have a single copy, and most of it doesn't undergo any genetic shuffling (a small region can recombine with the X chromosome). As a result, any mutations that occur on a single Y chromosome will always be inherited together. This makes the Y great for reconstructing history.
Let's say that, deep in our past, a mutation we call A occurred and gradually expanded in the population. Later, one of the descendants who carried A experienced a second mutation, B, which also expanded a bit. If we sequenced a population of 100 today, we might see 50 people who carry A and 30 who carry B. But every single person with B would also carry A. This idea allows us to infer the order in which these changes occurred and, given the average rate at which mutations appear, their timing.
Now, layer a bit of history on top of that. If A occurred after humans migrated out of Africa, it might be widespread in populations elsewhere around the globe. But if B occurred later, after further migrations, it might only show up in a specific region—say, Australia. So we can not only learn about the timing of different mutations, we can often figure out where they must have occurred.
But wait, there's more. If a population is relatively steady, there won't be much change in the number of mutations—for each new one that appears, another is likely to die out. But when a population is expanding, new mutations are more likely to be retained and show up in modern lineages. When making a Y chromosome family tree, this process will appear as a sudden burst of branches in a short amount of time.
Snails Reveal Ancient Human Migrations
In both places, the snails have large, white-lipped shells. And, according to the new work, the two groups also share genetic markers that are extremely rare elsewhere in Europe.
Along with other evidence, the findings offer a new window into ancient human migrations.
"It's interesting to use snail genetics to find out how snails colonize, and it also maybe gives us a little insight into what humans were doing, too," said Angus Davison, an evolutionary geneticist at the University of Nottingham in the United Kingdom.
"One really neat thing about this study is that, if we accept that humans transported snails, it really gives us a unique insight into an individual journey 8,000 years ago, and it gives us evidence of that from a source you might not imagine."
For more than 150 years, biologists have been puzzling over an Irish mystery: A number of wildlife species that live in Ireland are absent from the rest of Britain but are found in Iberia, the peninsula that includes modern-day Spain, Portugal and parts of France.
Research into this so-called "Irish question" has failed to produce a single theory that explains how and when various species covered hundreds of miles from one place to the other.
To see if they could add any new understanding to the Irish question, Davison and colleague Adele Grindon focused on a distinctive-looking snail that had the same one-inch long shells in both locations. According to fossil evidence, the snails first showed up in Ireland about 8,000 years ago. The mollusks had lived in southern Europe for tens of thousands of years before that.
First, the researchers enlisted volunteers to help collect nearly 900 snails from both parts of its range. Then, they extracted mitochondrial DNA, which is passed directly from mother to offspring, and they looked at specific areas of the genome that are known to vary from snail species to snail species.
The genetic material they analyzed was essentially identical between the two regional groups, the researchers report today in the journal PLOS ONE. The team was also surprised to find that they could trace the Irish population of snails directly back to a population in a specific region of the Pyrenees. The species lives nowhere in between.
The findings argue against a gradual move from one place to another, Davison said, and instead suggest that the snails migrated from Spain to Ireland in one step. He thinks it unlikely that birds transported the mollusks, partly because there are no known birds that migrate along that route that would have been large enough to carry the snails.
More likely, Davison suspects, people brought the snails with them as they moved. Ancient people may have intentionally brought the snails as a source of food on the trip. Alternatively, the snails may have hitched a ride in the grassy fodder packed for other animals.
In the early 2000s, some studies proposed a connection between Ireland and the Pyrenees, but those studies were later shown to be too small and flawed to be convincing, said Allan McDevitt, a geneticist at University College Dublin, who specializes on questions about the colonization of Ireland. The new research is far more robust.
"I think this study is important in that it does conclusively show that there is some link with Spain based on this snail," McDevitt said. "It's the most definitive proof yet that this is actually a very real migration that was happening."
Still unclear is how people made the trip. At least some of the journey would have been by sea, as Ireland was separated from the mainland by 15,000 years ago. But evidence of primitive canoe-like boats remains scant. Using new genetic tools to investigate unlikely creatures may be key to uncovering a better understanding of the past.
"Ireland has always been a very controversial topic, both in how people reached it there and how animals reached it there," McDevitt said. "We're finding a lot of things by looking at small animals. They actually do tell us a lot about how humans were moving."
Estimating pairwise coalescence rates with MSMC2 and fitting an IM model
To model the ancestral relationship between a pair of populations, we developed an isolation-migration model with a time-dependent migration rate between a pair of populations, which we call MSMC-IM. The approach requires time-dependent estimates of pairwise coalescence rates within and across two populations. To estimate these rates, we use an extension of MSMC , called MSMC2, which was first introduced in Malaspinas et al. 2016  (Fig 1A, Methods). MSMC2 offers two key advantages over MSMC . First, the pairwise coalescence model in MSMC2 is exact within the SMC’ framework , whereas MSMC’s model uses approximations that cause biases in rate estimates for larger number of haplotypes (S1 Fig). Second, since MSMC2 uses the pairwise tMRCA distribution instead of the first tMRCA distribution, it estimates coalescence rates within the entire range of coalescence events between multiple haplotypes, which ultimately increases resolution not just in recent times but also in the deep past. These two improvements are crucial for our new method MSMC-IM, which relies on unbiased coalescence rate estimates within and across populations, in particular in the deep past. Specifically, MSMC2 recovers simulated population size histories (with human-like parameters) well up to 3 million years ago, while keeping the same high resolution in recent times as MSMC (S1 Fig).
(A) MSMC2 analyses patterns of mutations between pairs of haplotypes to estimate local coalescence times along the genome. (B) MSMC-IM fits an isolation-migration model to the pairwise coalescence rate estimates, with time-dependent population sizes and migration rate. (C) As a result, we obtain the migration rate over time, m(t), and the cumulative migration probability, M(t), which denotes the probability for lineages to have merged by the time t and which we use to estimate fractions of ancestry contributed by lineages diverged deeper than time t.
Given MSMC2’s estimates of time-dependent coalescence rates within populations, λ11(t) and λ22(t), and across populations, λ12(t), we use MSMC-IM to fit an Isolation-Migration (IM) model to those three coalescence rates (see Methods). MSMC-IM’s model assumes two populations, each with its own population size N1(t) and N2(t), and a piecewise-constant symmetric migration rate m(t) between the two populations (Fig 1B, see Methods and S1 Text for details). Expressing the separation history between two populations in terms of a variable migration rate instead of the more heuristic relative cross coalescence rate facilitates interpretation, while maintaining the freedom to analyze data without having to specify an explicit model of splits and subsequent gene flow. Of the new parameters, the time-dependent migration rate m(t) is arguably the most interesting one, and it can be visualized in two ways (Fig 1C). First, the rates themselves through time visualize the timing and dynamics of separation processes, and second, the cumulative migration probability M(t) defined as which can be understood as the proportion of ancestry that has already merged at time t, and which makes it possible to quantify proportions of gene flow or archaic ancestry through time, as illustrated below. Being by definition monotonically increasing and bounded between 0 and 1, M(t) also turns out to be numerically close to the relative cross coalescence rate from MSMC . When M(t) becomes very close to 1, it means that lineages between the two extant populations have completely mixed into essentially one population. As a technical caveat, this means that at that time point our three-parameter model is overspecified. To avoid overfitting, we therefore employ regularization on m(t) and the difference of the two population sizes (see Methods).
Evaluating MSMC-IM with simulated data
We illustrate MSMC-IM by applying it to several series of simulated scenarios of population separation (see Methods). First, the clean-split scenario consists of an ancestral population that splits into two subpopulations at time T (Fig 2A). Second, the split-with-migration scenario adds an additional phase of bidirectional gene flow between the populations after they have split (Fig 2B). Third, the split-with-archaic-admixture scenario involves no post-split gene flow, but contains additional admixture into one of the two extant populations from an unsampled “ghost” population, which splits from the ancestral population (Fig 2C) at time Ta>T. In addition, to understand how MSMC-IM behaves under asymmetric demographic histories in the two populations, we consider the archaic-admixture-with-bottleneck-scenario (see Fig 2D). For each scenario, we simulated 8 haplotypes (four from each population), used human-like evolutionary parameters and varied one key parameter to create a series of related scenarios (see Methods). As discussed further below, to test internal consistency, we confirmed that MSMC-IM is able to infer back its own model, using simulations based on some of the genomic inferences carried out below.
(A) Clean-split scenario: Two populations with constant size 20,000 each diverged at split time T in the past, varying from 15kya to 150kya. (B) Split-with-migration scenario. Similar to A), with T varying between 15-150kya, and a post-split time period of symmetric migration (amounting to a total migration rate of 0.5 in both directions) between 10 and 15kya. (C) Split-with-archaic-admixture scenario: Similar to A), with T = 75kya, and population 1 receiving an admixture pulse at 30kya from an unsampled population that separates from the ancestral population at 1 million years ago. The admixture rate varies from 0% to 20%. (D) Split-with-archaic-admixture&bottleneck scenario: Similar to C), but with an added population bottleneck with factor 30 in population 1 between 40-60kya. Solid red lines indicate split times in all panels. In all plots, the blue light blue shading indicates the interval between 1–99% of the cumulative migration probability, the dark blue shading from 25–75%, and the black dashed vertical line indicates the median.
In the clean-split scenario, we find that MSMC-IM’s inferred migration rate m(t) displays a single pulse of migration around the simulated split time T (Fig 2A). This is expected, since in our parametrization, a population split corresponds to an instantaneous migration of lineages into one population at time T, thereby resulting in a single pulse of migration. In the split-with-migration series, we expect two instead of one pulse of migration: one at time T, as above, and a second more recent one around the time of post-split migration. In cases where the split time and migration phase are separated by more than around 20,000 years, this is indeed what we see (Fig 2B), although with some noise around this basic pattern. For less time of separation of the two migration pulses, MSMC-IM is not able to separate them in this scenario.
We also find two phases of migration for the split-with-archaic-admixture scenario, but this time with one phase around time T, and another one around the time of divergence of the archaic population Ta (Fig 2C). To understand this, consider how lineages in the two extant populations merge into each other (Fig 3B). One fraction 1-ɑ will merge into each other at the population split time T, as in the clean-split scenario. The other fraction, ɑ, will merge back only at the deep divergence time of the archaic lineage. These two merge events correspond to the two pulses we observe in Fig 2C—one at T and the other at the divergence time with the archaic population, Ta. Note that unlike in the above split-with-migration case, here there is no signal at the time of introgression, but only at the two split times. Inferring these two migration pulses in the presence of archaic admixture is robust to demographic events, as we show with the archaic-admixture-with-bottleneck scenario (Fig 2D), in which we introduced a bottleneck in one of the two extant population branches, similar in strength to the one observed in Non-African populations around 60 thousand years ago (kya) . We find, however, that in the presence of a bottleneck the second pulse is a bit more recent than expected (here at 1 million years ago).
(A) The cumulative migration probability M(t) is shown for selected simulation scenarios described in Fig 2B, 2C and 2D. Plateaus of M(t) indicate periods of isolation, with the level of the plateau indicating how much ancestry has merged before. (B) Schematic coalescence in the Split-with-archaic-admixture scenario. In this scenario, a fraction 1-α of lineages sampled from the two extant populations merges at time T, and the rest, of proportion α merges as time Ta. (C) For the split-with-archaic-admixture scenarios (with and without bottleneck), we can use the level of the plateau in M(t) to estimate 1-α, and thus α. The level of the plateau is measured at time t = 300kya.
We can analyze these multiple phases of migration in a more quantitative way, by using the cumulative migration probability, M(t), as introduced above. M(t) monotonically increases from 0 to 1 in all scenarios, exhibiting plateaus with gradient zero at times of no migration, and positive gradients in periods of migration (Fig 3A and S2 Fig). The level of these plateaus is indicative of how much ancestry has already merged at this point in time. Consider first the split-with-migration series (Fig 3A top panel), for which M(t) exhibits a plateau between the two migration pulses, at a level that corresponds to the amount of ancestry that has merged through the migration event. For this scenario, based on the simulated post-split migration rate between the two populations, we expect this plateau to be at around 0.64 (following the calculation in formula (64) in S1 Text). We find it to be higher than that, around 0.75, which we discuss further below. Consider now a scenario with archaic admixture (Figs 2C, 2D and 3A middle and bottom panels). At time T, at which both extant populations merge into each other, the cumulative migration probability reaches a plateau at a level around 1-ɑ, reflecting the fact that a proportion ɑ has not yet merged at point T, but is separated by a deeply diverged population branch. Only at time Ta, this branch itself merges into the trunk of the extant populations, thereby increasing M(t) from 1-ɑ all the way to 1. Based on this rationale, we can use visible plateaus in M(t) to estimate fractions of archaic or otherwise deep ancestry. Indeed, this rationale leads to estimates of archaic admixture proportions in our simulations which are accurate and robust to bottlenecks for rates of ɑ up to about 20%. For larger introgression rates, we find our estimates to be slightly underestimated. We attribute this to MSMC’s tendency to “overshoot” changes in coalescence rates, as can be seen in the relative cross-coalescence rates for larger values of alpha (S2C and S2D Fig), which causes the level of the plateau in M(t) to be higher than 1-ɑ, and hence ɑ to be underestimated. This is also the reason for the above-mentioned overestimation of the plateau in the split-with-migration scenario (Fig 3A top panel). This effect is more severe in the presence of a bottleneck (Fig 3C, blue curve) than without a bottleneck. Importantly, though, we find no evidence that M(t) exhibits plateaus below 1 in the absence of true deep ancestry, so this method can be considered conservative for detecting deep ancestry.
MSMC-IM also fits population sizes, which can be compared to the raw estimates from MSMC, i.e. to the inverse coalescence rates within population 1 and 2, respectively (see S1 Text for some non-trivial details on this comparison). We find that estimates for N1(t) and N2(t) are in fact close to the inverse coalescence rates, with some deviations seen in deep times, and in cases of archaic admixture. The latter is expected, given that estimated coalescence rates from MSMC2 capture both population size changes and migration processes, while in MSMC-IM these two effects are separated (S3 Fig).
Deep ancestry in Africa
We applied our model to 30 high coverage genomes from 15 world-wide populations from the SGDP dataset  (S1 Table) to analyze global divergence processes in the human past (Figs 4–6). When analyzing the resulting pairwise migration rate profiles, we find that several population pairs involving African populations exhibit by far the oldest population structure observed in all pairwise analyses. We find that in all population pairs involving either San or Mbuti, the main separation process from other populations dates to between 60-400kya, depending on the exact pair of populations (see below), but with small amounts reaching back to beyond a million years ago, as seen by the non-zero migration rates around that time (Fig 4A, S4 Fig), and the cumulative migration probability, M(t), (Fig 4B) which has not fully reached 1 until beyond a million years ago. Following the interpretation of M(t) as discussed above with the archaic-admixture simulation scenario, we can infer that in pairs involving San or Mbuti, at least around 1% of ancestry can be attributed to lineages of ancestry that have diverged from the main human lineage beyond 1 million years ago (see also Fig 7, discussed further below). The genetic separation profile in pairs involving Mbuti and San is, beyond the extraordinary time depth, not compatible with clean population splits (as seen in simulations, Fig 2A) or simple scenarios of archaic admixture, but instead shows evidence for multiple or ongoing periods of gene flow between (unsampled) populations. Between Mbuti and other African populations except San, we find three distinct phases of gene flow. The first peaks around 15kya, compatible with relatively recent admixture between Mbuti and other African populations. The second phase spans from 60 to 300kya, reflecting the main genetic separation process, which itself looks complex and exhibits two peaks around 80-200kya thousand years ago. The third and final phase, including a few percent of lineages from around 600kya to 2 million years ago, likely reflects admixture between populations that diverged from each other at least 600kya. In pairs that include San, the onset of gene flow with other populations is more ancient than with Mbuti, beginning at around 40kya and spanning until around 400kya in the main phase, and then exhibiting a similarly deep phase as seen in Mbuti between 600kya and 2 million years ago. We confirm that this deep divergence is robust to phasing strategy (see below) and filtering (see Methods). We also replicated this signal using an independent dataset  (S5 Fig). An exception to these signals seen with San and Mbuti are pairs involving Karitiana, which do not exhibit such deep divergence. This is likely due to the strong genetic drift present in Karitiana, and the low heterozygosity in that population , which may shadow deep signals.
(A) Migration rates. Dashed lines indicate the time point where 50% of ancestry has merged, and shading indicates the 1%, 25%, 75% and 99% percentiles of the cumulative migration probability (see Fig 2). (B) Cumulative migration probabilities M(t). Dashed lines indicate the relative cross coalescence rate obtained from MSMC2, for comparison. See S4 Fig for the full set of figures.
(A) Migration rates. Dashed lines indicate the time point where 50% of ancestry has merged, and shading indicates the 1%, 25%, 75% and 99% percentiles of the cumulative migration probability (see Fig 2). (B) Cumulative migration probabilities M(t). Dashed lines indicate the relative cross coalescence rate obtained from MSMC2. See S4 Fig for the full set of figures.
(A) Migration rates. Dashed lines indicate the time point where 50% of ancestry has merged, and shading indicates the 1%, 25%, 75% and 99% percentiles of the cumulative migration probability (see panel B). (B) Cumulative migration probabilities M(t). Dashed lines indicate the relative cross coalescence rate obtained from MSMC2. See S4 Fig for the full set of figures.
(A) Boxes show the 25% to 75% quantiles of the cumulative migration probability M(t), with bi-directional elongated error bars representing 1% and 99% percentiles. Colorcode: Red for African/African, blue for African/Non-African and orange for Non-African/Non-African pairs. (B) Barchart showing the amount of ancestry due to lineages older than 300, 600, 800kya and 1 million years ago, based on the cumulative migration probability M(t).
Apart from the deep structure seen with Mbuti and San, we find the second-most deep divergences between the West African Yoruba, Mandenka and Mende on the one hand, and French on the other (Fig 5A, S4 Fig, Fig 7 discussed further below), based on the time when M(t) reaches 99%. This might be consistent with recent findings of archaic ancestry in West-Africans [15,16], although it is not clear why the signal is primarily seen with French, and less consistently with Asian populations (Yoruba/Han have deep divergences, as well as Mende/Dai and Mandenka/Dai, but not other West-African/Asian combinations). Finally, pairwise analyses among Mende, Mandenka and Yoruba (Fig 4A, S4C, S4E and S4F Fig) exhibit a very recent migration profile, which appears to span up to about 20kya but not older, which is at odds with a recent finding of basal African ancestry present to different degrees in Mende and Yoruba . However, that signal may be too weak to be detected in our method, which is based on only two individuals per population.
Complex divergence between African and Non-African populations
Compared to the separation profiles between San or Mbuti and other populations, separations between other Africans and non-Africans look relatively similar to each other, with a main separation phase between 40 and 150kya, and a separate peak between 400 and 600kya (Fig 5 and S4 Fig). The first, more recent, phase plausibly reflects the main separation of Non-African lineages from African lineages predating the “out-of-Africa” migration event, and coinciding with the major population size bottleneck observed here (S6 Fig) and previously [3,4] around that time period. Signals more recent than about 60kya likely reflect the typical noisy spread of MSMC-estimated coalescence rate changes observed previously . The second peak of migration, between 400 and 600kya likely reflects Neandertal and/or Denisovan introgression into non-Africans. The age of that peak appears slightly more recent than, although overlapping with, previous split time estimates of those two Archaic groups from the main human lineage at 550-765kya . However, our simulation with archaic admixture with bottleneck (Fig 2D), shows that our model tends to underestimate the archaic split time in the presence of population bottlenecks as is the case for non-African populations [18–20]. In favor of the hypothesis that this second peak is caused by archaic lineages that have contributed to non-Africans is the fact that in all pairs of Papuans/Australians vs. Yoruba/Mende/Mandenka or Dinka, the second peak is particularly pronounced. This fits the archaic contribution hypothesis, since Papuans and Australians are known to have among all extant human populations the highest total amount of ancestry related to Neanderthals and Denisovans.
We investigated previous observations of potential ancestry from an earlier dispersal out of Africa, present in Papuan and Australian genomes [12,13,21]. Previously, one line of evidence for such a signal was based on shifts of relative cross coalescence rate curves between some Africans and Papuans or Australians on the one hand compared to curves with Europeans or East Asians on the other. With MSMC-IM we can compare these curves more quantitatively. While we were able to replicate this slight shift of relative cross coalescence rate or M(t) midpoint-based split times from African/Eurasian pairs to African/Australasian pairs reported in Ref.  using MSMC and Ref.  using MSMC2, we find that the estimated migration profiles of these pairs are very similar (S7 Fig), with a main separation midpoint around 70kya and a second older signal beyond 200kya, consistent with both Australasians and other Non-Africans being derived from a single genetic ancestral population without a more basal contribution to Australasians [12,13]. We conclude that the shift in the relative cross coalescence rate curve appears to be consistent with being caused by the higher amount of archaic ancestry present in Papuans and Australians. We note, however, that different separation events are not distinguishable in MSMC-IM when they are temporally close to each other, as we saw in the split-with-migration-scenario (Fig 2B).
Separations outside of Africa
All separations outside of Africa are younger than separations between Africans and Non-Africans, as expected (Fig 6, S4 Fig). The deepest splits outside of Africa are seen in pairs of Papuans or Australians with other Eurasians, in which the first peak of migration is seen at 34kya, corresponding to the early separation of these populations’ ancestors from other non-African populations after the out of Africa dispersal. In these pairs we see a second peak around 300kya, likely corresponding to the known Denisovan admixture in Papuans and Australians [13,22]. This is too recent for divergence time estimates between Denisovans and modern humans , which again is consistent with the underestimate seen in simulations with bottlenecks. Surprisingly, we see a similar second peak between French and Han, which is consistent with cross-coalescence rate features in previous observations [4,12] but of unclear cause. Consistent with the hypothesis that the second peak seen in Australasian/Eurasian pairs corresponds to Denisovan admixture, we do not see a second peak in the migration profile between Papuans and Australians, confirming that the gene flow likely occurred into the common ancestor of Australians and Papuans . The migration profile between Papuans and Australians shows a main separation between 15-35kya.
The second deepest splits in Non-African populations are seen between East Asian and European populations, which occur mostly between 20 and 60kya (cumulative migration probability midpoint at 34kya), followed by separations between Asian and Native American populations, between 20 and 40kya (midpoint at 28kya). The latter likely also reflects Ancestral North Eurasian ancestry in Native Americans , which is more closely related to Europeans than to East Asians, thereby pushing back the separation seen between East Asians and Native Americans. Finally, the most recent splits are seen between populations from the same continent: Dai/Han split around 9-15kya (midpoint 11kya), French/Sardinian around 7-13kya (midpoint 9kya) and within Native Americans around 7-13kya (midpoint 10kya) (Fig 6, S4 Fig).
To visualize the depth of ancestry in each population pair, we summarized all pairwise analyses by percentiles of the cumulative migration probability M(t) (Fig 7). Largely, Non-African pairs (orange) have their main separation phase, with the cumulative migration probability between 25% and 75%, between 20 and 60kya, with some more recently diverged pairs within continents. In contrast, African pairs (red) have their main phase largely between 60 and 200kya, with some notable exceptions of more recently diverged populations, and with the notable tail (99% percentile) up to 1 million years and older. Between Africans and Non-Africans, divergence main phases are largely within a similar window of 60-200kya as in African pairs, with three notable groups: divergence of Non-Africans from San falls between 80-250kya, from Mbuti between 70-200kya, and from other Africans between 50-150kya. To highlight the amount of ancestry contributed asymmetrically to one of the two populations from unsampled populations that diverged from the human lineage in the deep past (so-called archaic lineages), we show the distance of the cumulative migration probability from 1, 1-M(t), at different deep time points (Fig 7B). As described above, the deepest signals are seen in pairs involving San or Mbuti, reaching 3% of ancestry contributed from lineages that diverged at least 800kya, and around 1% of ancestry from lineages that diverged at least 1 million years ago. Similarly deep levels are seen in specific pairs involving French, in combination with the West African Mende, Mandenka and Yoruba and the East African Dinka, and for pairs Mende/Dai and Mandenka/Han, as discussed above.
Robustness to phasing and processing artifacts
MSMC2 (like MSMC) requires phased genomes for cross coalescence rate estimation, and we therefore rely on statistical phasing within the SGDP dataset, for which different strategies are possible. To compare the effect of selecting such phasing strategy, we generated phased datasets using eight different phasing strategies with three phasing algorithms (SHAPEIT , BEAGLE , EAGLE ). We included genotype calls from 12 individuals with previously published physically phased genomes  and then used those genomes to estimate the haplotype switch error rate. Among eight phasing strategies, SHAPEIT2 , without the use of a reference panel, but including information from phase-informative reads , resulted in the lowest switch error rate per kb (and per heterozygous site S8 Fig). Overall, switch error rates are higher in African populations, likely due to lower linkage disequilibrium, higher heterozygosity and relatively limited representation in the SGDP. To test how sensitive MSMC-IM is to different phasing strategies, we tested four phasing strategies on four different pairs of populations with evidence for extremely deep ancestry (Methods). We find that the migration profile from MSMC-IM is very similar for different phasing strategies. In particular, we find that the very deep signal seen in population pairs involving San and Mbuti is reproduced with different phasing strategies with and without a reference panel (S9A Fig). In a similar way, we confirmed the robustness of that signal with respect to choosing different filter levels (S9B Fig) and with respect to removing CpG sites, which are known to have elevated mutation rates (S9C Fig). We also explored to what extent switch errors affect our estimates using simulated data (S10 Fig), and confirmed robustness with respect to variation in recombination rates, which are assumed to be constant along the genome within MSMC2 but vary in reality (S11 Fig). Finally, to test internal consistency, we tested how well MSMC-IM was able to infer back its own model. We used the estimated migration rates and population sizes from eight population pairs (see Methods), and simulated genomic data under their inferred models. As shown in S12 Fig, the estimated migration patterns from the simulated and the real data are indeed very similar, including the deep signals seen in pairs with San and Mbuti.
Given the superiority of the read-aware phasing strategy with SHAPEIT without a reference panel [27,28] (S8 Fig), we used this method in all of our main analyses. However, even with this phasing strategy, the switch error rate is high in populations that are not well represented in the dataset. In case of indigenous Australians, the phasing quality is among the worst in the dataset (S8 Fig), arguably because the SGDP dataset contains only two Australian individuals (compared for example to 15 Papuans). To improve phasing in Australians specifically, we generated new high coverage genomic data for one of the two Australians in the SGDP dataset using a new library with longer read-pair insert sizes (see Methods). Using these additional reads reduced the switch error rate from 0.038/kb to 0.032/kb. (S8 Fig, blue isolated dot for Australian3). We ran MSMC2 on the long-insert Australian data, as well as the standard phased data, combined with one diploid genome from each of the other world-wide populations analyzed in this study. The inferred migration profiles from MSMC-IM (S13 Fig) for Non-African population pairs involving the long-insert phased Australian genome do not seem to be affected by the phasing method (S13 Fig). The migration profile from pairs of Africans versus the long-insert phased Australian tend to be slightly younger, but also show deeper structure in Dinka/Australian, compared to the same pair using the shapeit_pir phasing method, which uses phase-informative sequencing reads to improve phasing accuracy (Methods). Note that these migration rate densities exhibit more noise than the ones used in our main analysis (S4L Fig), since they are based on only one individual per population, while the main analyses are based on two individuals per population. The main separation between Papuan and Australian remains at 15-35kya, as shown in the migration profile from both phasing strategies, very close to the estimates from 8 haplotypes in the main analysis (S4L Fig), and earlier than the previous estimates of 25-40kya .
Similar to the procedure introduced for PSMC , we use a block-bootstrap approach to assess statistical uncertainty of our method. We find that there is very little uncertainty around MSMC-IM’s migration rate estimates (S14 Fig) based on these bootstrap-estimates. This should be taken with caution, though, since the bootstrap is only able to address the uncertainty caused by randomness in the data, not by systematic biases. We know that MSMC typically “smears out” sudden changes in coalescence rates, which is due to the wide variance in local estimates of coalescence times, and this type of error is not revealed by the bootstrap. It does, however, give high confidence to specific results, such as estimates of archaic ancestry between 1 and 20% as seen in Fig 3C. According to our bootstrap test (S14 Fig), the cumulative migration probability M(t) does hardly vary at all in bootstrap replicates, so estimates of deep ancestry fractions such as the ones shown in Fig 3C and Fig 6B for real data, are very accurate.
Organizations Against Human Trafficking
A great help to governments all over the world are local an international Non-Governmental Organizations (NGOs) who actively assist governments in combating human trafficking. A number of them were created exclusively to fight human trafficking, such as Called to Rescue, Polaris, and Anti-Slavery International. Other NGOs participate and cooperate against human trafficking, such as Save the Children, ChildHope, Women’s Rights Worldwide, and Amnesty International.
The fight against human trafficking may be joined by donating to NGOs, volunteering to work with local NGOs, and reporting suspicions of human trafficking rings. Furthermore, achievable for all citizens is awareness of and self-protection against human trafficking schemes through responsible travel, self-defense, and caution in any time of recruitment or deal-making.
Live bird migration maps
Real-time analysis maps show intensities of actual nocturnal bird migration as detected by the US weather surveillance radar network between local sunset to sunrise. All graphics are relative to the Eastern time zone. When present, the red line moving east to west represents the timing of local sunset, the yellow line represents the timing of local sunrise. Areas with lighter colors experienced more intense bird migration. Orange arrows show directions to which birds flew. Green dots represent radar locations for which data are available red dots represent radar locations with no data available. Note that many radars in mountainous areas (e.g. the Rockies) have obstructions that restrict radar coverage, providing the appearance of no migration where migration may be occurring.
Brighter colors indicates a higher migration traffic rate (MTR) expressed in units birds/km/hour. The migration traffic rate indicates the number of birds per hour that fly across a one kilometer line transect on the earth’s surface oriented perpendicular to the direction of movement of the birds.
Cornell Lab of Ornithology currently produces these maps. Support for this research came from NASA, Edward W. Rose Postdoctoral Fellowship, and Amazon Web Services. The BirdCast project was created by grants from the National Science Foundation and supported by additional grants from Leon Levy Foundation.
View more than 150 years of hurricane tracking data in your region. Shown here: Category 4 and 5 hurricane tracks that crossed over the state of Florida between 1910 and 2018.
NOAA's Historical Hurricane Tracks is a free online tool that allows users to track the paths of historic hurricanes. The site, developed by the NOAA Office for Coastal Management in partnership with NOAA's National Hurricane Center and National Centers for Environmental Information, offers data and information on coastal county hurricane strikes through 2016. It also provides links to detailed reports on the life histories and effects of U.S. tropical cyclones since 1958, with additional U.S. storm paths traced as far back as 1851. The site also contains global hurricane data from as far back as 1851.
Historical Hurricane Tracks allows users to search by place name, storm name or year, or latitude and longitude points. With the search results, users can generate a map showing the track of the storm or storms accompanied by a table of related information.
Nature’s Most Impressive Animal Migrations
Migration is a natural phenomenon that has fascinated humans for centuries and is important to ecosystem health.
Biology, Ecology, Geography
Arctic Tern (Sterna paradisaea)
Arctic Tern (Sterna paradisaea) flying over Rudolf Island, Russia. One of the many stops on their long journey.
Photograph by: Cory Richards
Migration is a natural phenomenon observed in species across the animal kingdom, from the tiniest insects to the gargantuan blue whale ( Balaenoptera musculus) . Every year, millions of animals set out on an epic journey in search of food, shelter, and mating opportunities. Often travelling thousands of miles by land, sea, or air, these animals push the limits of endurance.
Arctic terns ( Sterna paradisaea) are small, plain-looking birds, weighing between 90&ndash120 grams (3.2&ndash4.2 ounces) with a wing span of 64&ndash76 centimeters (25.2&ndash29.9 inches). To the untrained eye, they do not look as if they are built for endurance, but these birds take the trophy for the longest migration of any animal in the world.
Flying from pole to pole, Arctic terns spend most of their year at sea chasing a perpetual summer. Seasons are reversed in the Northern and Southern hemispheres, so as winter approaches in their Arctic breeding grounds, the terns head south to the Antarctic where summer is just beginning. Arctic terns are believed to migrate around 40,000 kilometers (25,000 miles) a year, but a recent scientific study suggests that they might fly double that distance.
Perhaps one of the most famous migrations is the multi-generational round trip of the monarch butterfly. Monarch butterflies ( Danaus plexippus) can be found all over the United States and further afield, but it is the northeastern American population that is famous for making the 4,800-kilometer (3,000-mile) journey from Canada to Mexico.
Each year, millions of monarch butterflies leave their northern ranges and fly south to the oyamel fir forests near the Sierra Madre mountains, where they gather in huge roosts to survive the winter. When spring arrives, the monarchs start their return journey north the population cycles through three to five generations to reach their destination. Along the way, females lay eggs on milkweed plants, which the caterpillars use for food after hatching. This new generation of butterflies complete the journey their great-great-great-grandparents started. It is still a mystery to scientists how the new generations know where to go, but they appear to navigate using a combination of the Earth&rsquos magnetic field and the position of the sun.
The Great Migration
Wildebeest ( Connochaetes taurinus) take the crown for the most dramatic migration. Wildebeest (also known as gnu) are members of the antelope family, but they look more like cows with their big horns, stocky build, and shaggy manes. They live in huge groups of over one million individuals, along with thousands of zebras (Equus quagga) and gazelles ( Eudorcas thomsonii) . During the dry season, this giant herd roams the Serengeti-Mara ecosystem of Tanzania and Kenya in search of fresh grass and water. It is a round-trip that spans hundreds of miles and two countries. The herd moves as a great swarm, and individuals must keep up or risk being picked off by the lions (Pantera leo), hyenas (Crocuta crocuta), and crocodiles (Crocodylus niloticus) that gather to hunt.
Slow and Steady Wins the Race
Humpback whales ( Megaptera novaeangliae) are one of the largest animals on the planet, weighing in at an impressive 36,000 kilograms (79,366 pounds). Fully grown adults can reach 18 meters (59 feet) in length and can live over 48 years. These giants spend their summers at feeding grounds in cold, nutrient-rich waters that support an abundance of krill and small fish. In the winter, they migrate to warmer waters to raise their calves and avoid predation by killer whales.
It is a journey that can take over 8,000 kilometers (4,970 miles) each way, making it the longest migration of any mammal on Earth. Humpback whales are slow swimmers, but they make up for it by traveling non-stop for days at a time. They do not feed along their migration route and instead survive on fat reserves built up during the summer months.
Salmon ( Salmo salar) spend most of their lives in the Atlantic and Pacific oceans, where they feed and grow before migrating back to the rivers where they were born. Salmon swim across the ocean to the mouth of the river, navigating using a combination of chemical cues, the sun, and the Earth&rsquos magnetic field. To reach their final destination, the salmon must swim up the river, in an event known as &lsquothe salmon run&rsquo. In an incredible feat of endurance, they swim up to 400 kilometers (250 miles) against the current, battling rapids and leaping up waterfalls, all while avoiding predators that congregate along the banks in hopes of catching a nutritious meal. When they finally reach their birth place, the salmon spawn and then die.
Animal migration has fascinated humans for centuries. Scientists have shed light on some of the enduring mysteries about how species navigate and what drives them to leave a habitat, but we still have a lot to learn. These incredible journeys are certainly captivating, but they also have a vital role to play in the ecosystem. Migration affects the distribution of prey and predators, keeps nutrients cycling around the planet, helps with the spread of pollen and seeds, and even influences human economies. Animal migrations are impressive, but they are also essential for a healthy ecosystem and, ultimately, a healthy planet.
Arctic Tern (Sterna paradisaea) flying over Rudolf Island, Russia. One of the many stops on their long journey.
3 Case Studies
The report profiled three countries as case studies—Ethiopia, Bangladesh and Mexico—and warns that fast-growing cities will have to diversify economically and create climate-resilient jobs to successfully absorb population growth.
There are exceptions. Declining rainfall in Ethiopia’s northern highlands, for example, may drive people out of the country in search of new areas where they can grow rainfed crops. And lack of rainfall in Addis Ababa, Ethiopia’s largest city, may slow its growth.
Alternatively, sea-level rise and storm surges will prompt growth in the major cities of Bangladesh, including the capital city of Dhaka. Bangladesh, the study predicts, will experience greater shifts and changes to population from climate change than any other event.
Mexico, the wealthiest of the trio profiled, is less vulnerable to climate change and better prepared than Ethiopia and Bangladesh. But “it needs to pay close attention to pockets of poverty,” the study’s authors found. The central plateau around Mexico City and Guatemala City, which may offer better climate conditions, may attract climate migrants.
But there’s not a lot of time to act. Without cuts to greenhouse gases and other preparations, climate migration will most likely rise through 2050, the authors find, and then accelerate.