artificial intelligence

In this news article by Tina Hesman Saey, she talks about a new deep-learning artificial intelligence system called AlphaFold. The purpose of this new technology is to predict the 3-D shapes of proteins by recognizing patterns in structures. At its first release in 2021, the AlphaFold included predictions for most known human proteins, with predictions of over 350,000 protein structures. Since its initial release, the AlphaFold database has increased the number of protein structure predictions to over 200 million.

As you can see by the model, some parts of the prediction are more or less accurate than others. One of the main issues with AlphaFold is the fact that these are only predictions. Scientists cannot use this information with full confidence and require further experiments to be able to be confident in the findings. Even with this issue, scientists and researchers have been able to develop potential new vaccines, improve their understanding of diseases, and gain insight into human evolution with this new artificial intelligence system.

As you may remember from your AP Biology class, protein structures are extremely important in allowing the protein to perform its job. Proteins are extremely integral in your body’s ability to function. Enzymes are used to speed up reactions, antibodies protect our body from diseases, hormones send signals to the body, etc. By knowing the structure of the protein, scientists gain a better understanding of how proteins and all of these processes work. Just by changing one amino acid, the whole structure changes which can cause diseases such as Sickle Cell Anemia. If you would like to learn more about other important purposes of protein structure, this article goes into more in-depth on the applications of protein structure and modeling.

Now onto the topic of artificial intelligence as a whole. While artificial intelligence can reduce human error, take risks instead of humans, create unbiased decisions, and automate repetitive actions, it also has several downsides(besides the possibility of it becoming sentient and taking over the world like in movies). Some of these are having high costs, increasing the amount of unemployment, being emotionless, having no morals, and making humans lazier. Artificial intelligence has the potential to completely change society as a whole. I believe that as long as we are able to keep artificial intelligence under control and not let it get into the wrong hands, it will be a great benefit to society through important breakthroughs, such as AlphaFold, and new ideas that would have never been thought of without the use of this technology.

Do you believe artificial intelligence will be the salvation for humanity or will it be its downfall?

Customizing Cancer?

By clalvincycle

On February 16, 2022

In Student Post

Oncologists are moving toward a future in which cancer treatment is customizable and specific to each patient. This is achieved through genomic testing.

As genes differ from person to person, the information from genomic testing is unique to each. This speaks to what we’ve learned recently in bio class about how cellular mutations cause cancer. Changes to the DNA of a cell, specifically to genes that control the cell cycle, could result in oncogenes. But, I digress. The drawback of this type of treatment is that oncologists are met with so much information that it becomes not useful, making the treatments less personal than they should be.

CHALLENGING GENETICS

The reason for this is due to the inability to identify which test will be most useful for each patient. When genetic data is obtained, what it means for the patient isn’t exactly clear. The example given in an article by Andrew Ip states that “several inhibitors of the enzyme anaplastic lymphoma kinase (ALK) have proven effective in treating lymphoma, non-small cell lung cancer, and neuroblastoma, while other findings suggest one of the inhibitors can treat pediatric oncology patients”. Although information was found from research specifically for ALK, it appears that much more is affected by these inhibitors making it less “personalized” than intended.

Another reason as to why genomic testing is so difficult is accessibility. It can be difficult to get the treatments to patients due to health insurance limitations. In another instance, according to Andrew, “oncologists in community settings … had difficulty handling tumor samples, faced long turnaround times for laboratory tests, and had limited access to targeted therapies. To make it more difficult, next-generation sequencing results are often provided as a pdf, which cannot be digitally integrated with a patient’s electronic health records”. It appears altogether that oncologists are hindered by the lack of seamless integration of genomic testing into daily scenarios.

THERE IS HOPE

Although it appears that oncologists are overwhelmed, there is progress being made to support them.

At the Hackensack Meridian Health John Theurer Cancer Center, where Andrew practices oncology, genomic testing was put into action. An ill patient had two separate biopsies done, and the findings of both contrasted each other greatly. One specified that the cancer identified was incurable, while the genomic sequencing depicted the cancer as curable. The patient was treated with chemotherapy and made quick improvement.

The Genomic Testing Cooperative joined with Hackensack Meridian Health to implement an “in-house genomic profiling center”. As stated in Andrews article, the center “analyzes 434 genes for solid tumors, searching for DNA and RNA mutations and chromosomal structural abnormalities. For blood cancers, the service generates a 177 gene panel hematology profile”.

This isn’t all. A new database to which will aid oncologists in using the genomic results, cancer types, cancer medicines and patient outcomes is being built there as well.

FUTURE ADVANCES

In order to fully take advantage of genomic sequencing, companies are turning toward artificial intelligence. The goal is for AI to be able to use information from genomics, drug trials, patient demographics, and past scientific research to provide its own efficient course of action. This is called a clinical decision support system or CDS. IBM Watson was to be a CDS but did not suffice.

Until then oncologists take what Andrew describes as a “holistic approach to care”. This involves working with multidisciplinary teams made up of radiologists, pathologists, medical oncologists, radiation oncologists, and surgeons. Altogether they are known as molecular tumor boards. It’s fascinating to see just how much goes into making cancer care especially personalized to each patient.

A New Way Of Detecting the COVID-19 Virus

By addna

On December 20, 2021

In Student Post

In a study conceived by Mayo Clinic investigators it was found that Artificial Intelligence may offer a new way of detecting if a person has contracted Covid-19. Researchers found that the Covid-19 virus creates small electrical changes in the heart that Artificial Intelligence (AI) can detect and be used for a new form of a rapid, reliable test. Since Covid-19 has a 10-14 day incubation period, symptoms take long periods of time to show up. Once patients do show symptoms it is hard to access a reliable Covid test with fast results. An Artificial Intelligence enhanced EKG is a rapid and cost effective alternative for Covid testing.

This study was done on a racially diverse population of volunteers from 14 different countries. Patients selected had previous EKG data from when they were diagnosed with Covid-19. This data was compared with the EKG data of patients not infected by Covid-19. AI was then trained to detect the subtle changes in the heart by more than 26,000 EKG’s and tested on 7,800 EKG’s that were not previously used. The prevalence of Covid-19 was about 33% and the accuracy of the negative predictive value of the AI was about 99.2%.

For any form of Covid test, accuracy is the most important value. The study shows the consistency of biological signals in the EKG and the Covid-19 infection. To confirm that Artificial Intelligence will be a helpful factor in our fight against the pandemic, this study needs to be tested on asymptomatic people.

In AP biology class this year, my class has learned about sending signals between cells. A heartbeat happens when the SA node (pacemaker of the heart) sends out an electrical impulse.The upper chambers of the heart contract and the AV node sends an impulse into the ventricles.The lower heart chambers then contract and the cycle starts over again.

A Living Supercomputer is No Longer Science Fiction

By mahmmalia

On February 29, 2016

In Student Post

“A supercomputer that’s alive? No way, sounds like a bunch of nonsense sci-fi to me!” Well, an international team of researchers at McGill University would beg to differ.

These researchers recently created a supercomputer that uses biological agents as an energy source, rather than electricity. Their new “biocomputer” is also significantly smaller and more energy efficient than the typical supercomputer.

The secret? ATP, or energy used by almost all living, breathing creatures. ATP speeds up the “thinking” process in a computer by allowing it to make parallel computations instead of sequencing only one signal at a time.

(Source: https://en.wikipedia.org/wiki/Adenosine_triphosphate#/media/File:ATP-xtal-3D-balls.png)

Strings of biological agents arranged in a very miniscule (on the nano scale) but highly organized circuit grid are powered by ATP to move in a controlled fashion to send parallel signals in the computer.

One of the most famous supercomputers, IBM’s Watson, is very large, as depicted by the image below:

(Source: https://en.wikipedia.org/wiki/Watson_%28computer%29#/media/File:IBM_Watson.PNG)

But this new biocomputer is about the size of a book due to the nano scale of the biological agents. Also due to these agents, there is less heat produced, so the computer uses a lot less energy than an electronic supercomputer (which often needs its own power plant to function).

Of course, further research needs to be done, but this creation of the biocomputer is a huge achievement in the world of artificial intelligence.

Check out this other article regarding the new biocomputer, and these articles on more information about the supercomputers of today, such as Watson and Google’s AlphaGO, and artificial intelligence in general.

Questions? Thoughts? Opinions? Worries? Leave a comment below!

Original Article

BioQuakes

AP Biology class blog for discussing current research in Biology

Tag: artificial intelligence

AlphaFold: Is Artificial Intelligence Taking Over?

Customizing Cancer?

A New Way Of Detecting the COVID-19 Virus

A Living Supercomputer is No Longer Science Fiction

Archives