Machine learning will help reduce the tsetse fly population to reduce the incidence of sleeping sickness
Hi, Habr! I present to your attention the translation of the article " Machine learning can control tsetse flies and thus reduce sleeping sickness ".
Tsetse females mate once in their lives, which makes it possible to control the size of the population of these harmful insects. So a female who mates with a barren male will not have offspring. By controlling a sufficient number of matings, as a result, their population can be reduced, hence, the incidence of sleeping sickness among people and cattle can be reduced.
A study conducted in Senegal has shown that this idea is feasible. Over the past five years, tsetse males, sterilized using gamma rays, have been released three times a week to infected areas. This reduced the local population of flies by 98%, with a corresponding decrease in the incidence of sleeping sickness. But such projects require a huge number of sterile males, which need to be bred and delivered in a timely manner, which is difficult.
One of the problems is that breeding males inevitably affects breeding females. Sorting by gender is necessary in order to irradiate exclusively males. Elementary irradiation of both sexes causes problems, since the sterilization of females requires a higher dose of radiation, which can cause the death of males. Sorting Tsetse lies in waiting until the flies hatch from the pupae. At the same time cooling them to reduce metabolism and, consequently, their activity. The separation of males from females is done manually with a brush. The male differs from the female by the presence of antennae, which helps to identify it. This process is effective but time consuming and time consuming. Zelda Moran from Columbia University believes there is a better way.
In 2014, Ms. Moran, who at the time was a researcher in the entomology laboratory of the International Atomic Energy Agency, in Vienna, doing a similar job, noticed that the pupae of the female and the male tsetse develop differently. Adult flies emerge from the pupae 30 days after pupation. Although the tsetse fly pupae are opaque, Miss Moran found that under certain lighting conditions, such as infrared rays, it was possible to notice that the wings of the insects began to darken. In the case of females, this occurs approximately 25-26 days after pupation. In the case of males, this happens later: 27-29 days after pupation. In principle, this makes it possible to sort the flies before they come out of their pupae.
It was necessary to use this method until Ms. Moran accidentally met Dr. Szabolcz Mark, the astrophysicist of Colombia. At that time, Dr. Mark used machine learning to find patterns in large groups of astrophysical data. He proposed to apply a similar process to the sexual definition of pupae.
At first, Dr. Mark and his colleagues used an infrared scanner to create images of a large number of pupae. Then they used these images to teach a computer algorithm to decide whether the pupa is defined as male, female, or not yet formed. It also allows you to separate the dead from the living individuals. Machine learning made it possible to automatically sort the live pupae of males from the rest, with the help of air bubbles or water streams, in order to weed out unwanted individuals. Thus, individuals may be irradiated and released.
Machine learning should simplify male sterilization for projects such as the Senegal project. Perhaps it can be used in other forms. If the definition of sex by machine learning could be applied to other insects, such as mosquitoes, diseases such as malaria and dengue fever could also be controlled by humans.
Tsetse females mate once in their lives, which makes it possible to control the size of the population of these harmful insects. So a female who mates with a barren male will not have offspring. By controlling a sufficient number of matings, as a result, their population can be reduced, hence, the incidence of sleeping sickness among people and cattle can be reduced.
A study conducted in Senegal has shown that this idea is feasible. Over the past five years, tsetse males, sterilized using gamma rays, have been released three times a week to infected areas. This reduced the local population of flies by 98%, with a corresponding decrease in the incidence of sleeping sickness. But such projects require a huge number of sterile males, which need to be bred and delivered in a timely manner, which is difficult.
One of the problems is that breeding males inevitably affects breeding females. Sorting by gender is necessary in order to irradiate exclusively males. Elementary irradiation of both sexes causes problems, since the sterilization of females requires a higher dose of radiation, which can cause the death of males. Sorting Tsetse lies in waiting until the flies hatch from the pupae. At the same time cooling them to reduce metabolism and, consequently, their activity. The separation of males from females is done manually with a brush. The male differs from the female by the presence of antennae, which helps to identify it. This process is effective but time consuming and time consuming. Zelda Moran from Columbia University believes there is a better way.
In 2014, Ms. Moran, who at the time was a researcher in the entomology laboratory of the International Atomic Energy Agency, in Vienna, doing a similar job, noticed that the pupae of the female and the male tsetse develop differently. Adult flies emerge from the pupae 30 days after pupation. Although the tsetse fly pupae are opaque, Miss Moran found that under certain lighting conditions, such as infrared rays, it was possible to notice that the wings of the insects began to darken. In the case of females, this occurs approximately 25-26 days after pupation. In the case of males, this happens later: 27-29 days after pupation. In principle, this makes it possible to sort the flies before they come out of their pupae.
It was necessary to use this method until Ms. Moran accidentally met Dr. Szabolcz Mark, the astrophysicist of Colombia. At that time, Dr. Mark used machine learning to find patterns in large groups of astrophysical data. He proposed to apply a similar process to the sexual definition of pupae.
At first, Dr. Mark and his colleagues used an infrared scanner to create images of a large number of pupae. Then they used these images to teach a computer algorithm to decide whether the pupa is defined as male, female, or not yet formed. It also allows you to separate the dead from the living individuals. Machine learning made it possible to automatically sort the live pupae of males from the rest, with the help of air bubbles or water streams, in order to weed out unwanted individuals. Thus, individuals may be irradiated and released.
Machine learning should simplify male sterilization for projects such as the Senegal project. Perhaps it can be used in other forms. If the definition of sex by machine learning could be applied to other insects, such as mosquitoes, diseases such as malaria and dengue fever could also be controlled by humans.