Deep learning has grown exponentially over the last few years, marking its territory in several verticals. But now, this sought after the technology has started hunting for ghosts. Recently, a paper has been published in Nature Communication, showing a ghost population that has also contributed to today’s genomes.
Sign up for your weekly dose of what's up in emerging technology.
Talking about the unknown population, researchers have stated that it is either related to the Neanderthal-Denisova clade or diverged early from the Denisova lineage.
Also, the study shows how AI can be used in palaeontology — whether it is about identifying unforeseen ghosts or about uncovering the faded footprints of the evolutionary processes. According to the paper, researchers have built a deep learning-based demographic in an Approximate Bayesian Computation framework in order to know more about the evolutionary history of Eurasian populations and this also includes past introgression events in Out of Africa (OOA) populations fitting the current genetic evidence.
According to reports, researchers have generated numerous simulated evolutionary data such as the number of ancestral human populations, their sizes, when they diverged from one another, their rates of intermixing and so on. And from those data, a huge number of simulated genomes for present-day human beings have been generated.
This new deep learning method explains the level of gene flow that is too small for the usual statistical approaches. Scientists trained their deep learning algorithm on genomes so that it learned which kinds of evolutionary models were most likely to produce given genetic patterns.