Artificial Intelligence Study of Human Genome Finds Unknown Human Ance

Category: Entertainment

Presentation Description

No description available.


Presentation Transcript

slide 1:

9/23/2019 Artificial Intelligence Study of Human Genome Finds Unknown Human Ancestor 1/5 Articial Intelligence Study of Human Genome Finds Unknown Human Ancestor usm systems Sep 23 · 5 min read A recent study used machine learning technology to analyze eight leading models of human origins and evolution and the program identified evidence in the human genome of the “ghost population” of human ancestors. Analyzes indicate that a group of previously unknown and extinct hominins with Homo sapiens in Asia and Oceania interfered somewhere along the long winding road of human evolutionary history leaving only fragmented traces in modern human DNA.

slide 2:

9/23/2019 Artificial Intelligence Study of Human Genome Finds Unknown Human Ancestor 2/5 The study published in Nature Communications is the first example of how machine learning can help us uncover our own origins. By storing large amounts of genomic data left in fossil bones and comparing it with DNA in modern humans scientists can begin to fill in some of the gaps in the evolutionary history of our species. In this case the results appear to match the theories of paleoanthropology developed by studying human ancestral fossils found on Earth. New data suggest that the mysterious hominin came from the amalgamation of the Neanderthals and the Denisovans who were only identified as a distinct species in the human family tree in 2010. In our evolutionary past such a species looks like a fossilized 90000-year-old teenage girl from the Denisova cave of Siberia. Her remains were described last summer as a unique example of a first-generation hybrid between the two races with a Neanderthal mother and a Denisovan father. Jam Bertranspiett the co-author of the study of evolutionary biology at the University of Barcelona’s Pompey Fabra said “We are certainly a kind of person at the origin of this population but not just one person but the entire population.” Prevalence of the homo genus The ability of early humans to adjust to changing conditions eventually led to the initial species of Homo migrating living and spreading from Africa to Eurasia 1.85 million years ago. Image courtesy Anton Potts and Aiello 2014 Science 345 6192 Previous human genetic studies suggest that after modern humans left Africa perhaps 180000 years ago they later interfered with species such as the Neanderthals and the Denisovans who lived with early modern humans before they became extinct. But it is very difficult to re-draw our family tree to include these different branches. Evidence for “ghost” species is scant and there are many competing theories to explain when where and how often Homo sapiens interfere with other species. Traces of these ancient interspecies connections known as interrogations can be traced to differences in the human genome. If both chromosomes come from the same human species scientists will notice a greater separation between the two chromosomes than

slide 3:

9/23/2019 Artificial Intelligence Study of Human Genome Finds Unknown Human Ancestor 3/5 you might expect. When scientists sequenced the Neanderthal gene in 2010 they realized that some of these differences represent differences in our genes from the Neanderthal. Studies have shown that some living humans belong to 5 percent of their ancestors to Denisovans. “So we thought we were going to try to find these regions that are highly divergent in the genome see these as Neanderthals and Denisovans and then see if they explain the whole picture” Bertranpet said. “When this happens if you remove the Neanderthal and Denisovan parts the genome is still very different.” Identifying and analyzing many different regions across the genome and counting the countless gene combinations that can be produced is a huge task for humans to solve on their own — but it is a work for deep learning algorithms. Deep learning is a kind of artificial intelligence in which algorithms are designed to act as an artificial neural network or a program that can process information similar to the mammalian brain. These machine learning systems can identify patterns of “learning” previous information and enable them to perform new tasks or look for new information after analyzing enormous data. “Deep learning is a more complex shape for a set of points in a larger space” says Joshua Schreiber an evolutionary geneticist at Temple University. “Instead of placing a line between Y and X you have to fit some point in a very large thousand-dimensional space. Deep Learning says “I don’t know what shape fits these points but let’s see what happens.” In this case machines are set to analyze the human genome and assess the human population by simulating how our DNA has evolved in many thousands of scenes of ancient evolution. The program sought to combine the structure and evolution of DNA as well as the patterns of human migration and fertility into some of the most intricate pieces of the puzzle. Researchers trained the computer to analyze eight different models of the most plausible theories of early human evolution throughout Eurasia. These models come from previous studies which lead to the current picture of the human genome including its known Neanderthal and Denisovan components.

slide 4:

9/23/2019 Artificial Intelligence Study of Human Genome Finds Unknown Human Ancestor 4/5 “There may be other models but these models are being proposed by other people in the scientific literature” says Bertranspiett. Each model starts with an accepted Off-Africa event followed by different divisions between human populations including different breeds of known species and possible “ghost” species. The human family tree Humans or Homo sapiens come from a complex tree of upright ancestors including the genus Ardipithecus Australopithecus and Paranthropus. Smithsonian’s Human Origins Program “Of these eight models we calculate weekly counts of how well they can reach the actual and current genetic composition of humans” says Bertranpetit. “Every time we simulate it it’s a simulation of the possible path of human evolution and we’ve run those simulations thousands of times and deep learning algorithms can detect models that are better suited to the data.” The end of the machine There is an ancestral race in our lineage that we have not yet identified. “So far the only models we have tested that are really supported by the data are these ghost population interrogation” says Bertranpet. Despite the increasingly complex ancient world of Eurasia and Oceania intriguing study and others may help to re-map the map of how humans migrated and evolved. “This is certainly interesting and consistent with the image of complex reticulated phylogeny in human evolution” University of Pennsylvania Population Geneticist Ian Matheson said via email. “I’m not sure it makes sense to talk about ‘interrogation events’ when that seems ideal.” In fact only eight models have been tested and many more are possible. Mathison’s new findings are “perfectly acceptable but the reality is much more complicated.” As new fossil discoveries have been made in this field updated samples can now be tested on the human genome using these types of programs. Schreiber said the power of deep learning to study human origins certainly lies in its ability to analyze complex models.

slide 5:

9/23/2019 Artificial Intelligence Study of Human Genome Finds Unknown Human Ancestor 5/5 “If you’re an anthropologist and you want to make a very detailed model and you want to know if this interrogation happened 80000 years ago or 40000 years ago that’s the power of a deep learning approach like this.” In complex terms the fertility of ancient Eurasia is still only a part of our human story. Bertranpet thinks that future advances in deep learning will help uncover other new chapters. “This type of analysis method is going to get all kinds of new results” he says. “I know that people in Africa will find extinct groups that have not yet been identified. There is no doubt that Africa is going to show us some surprises in the future.” Want to know about Ai services then have a free visit for USM systems Science Machine Learning Articial Intelligence Ai Services Ai Solution About Help Legal

authorStream Live Help