A Brazilian study shows how wildfires and forest burning for agriculture influence rain cloud formation in the Amazon. According to the authors, aerosols (tiny solid particles and liquid droplets emitted into the atmosphere by fire) hinder the freezing of cloud droplets when the atmosphere is humidified, but can also promote freezing when the atmosphere is dry. This alters the natural functioning of clouds and their typical height, and may also affect precipitation and the amount of sunlight reaching the ground. 

To arrive at this conclusion, the scientists used a large dataset collected over 15 years, from 2000 to 2014, involving satellite imagery from the United States National Oceanic and Atmospheric Administration (NOAA), measurements of atmospheric aerosols from fires made by NASA’s Aerosol Robotic Network (AERONET), and reanalysis data from the European Center for Medium-Range Weather Forecasts (ECMWF). Reanalysis data provides the most complete picture currently possible of past weather and climate, blending observations and past forecasts rerun with modern forecasting models, according to ECMWF.

The satellite images and reanalysis data covered the entire Amazon region. The aerosol data referred to five locations in southern Amazonia: Alta Floresta and Cuiabá in Mato Grosso state; Rio Branco in Acre state; and Ji-Paraná and Ouro Preto do Oeste in Rondônia state.

The purpose of the investigation was to perform an observational study of the temperature at which droplets freeze in convective clouds, which form vertically and can reach heights exceeding 10 km, to identify the key drivers of the phenomenon. The presence of ice in clouds is important since it influences the formation of rain and the average time clouds remain in the atmosphere. “The longer clouds last on average, the more solar radiation is reflected back into space, contributing to the cooling of the planet,” said Alexandre Correia, a professor in the Department of Applied Physics at the University of São Paulo’s Institute of Physics (IF-USP) and first author of the article.

The study was supported by FAPESP. The co-authors were Elisa Sena (Federal University of São Paulo), Maria A. F. Silva Dias (Institute of Astronomy, Geophysics and Atmospheric Sciences, IAG-USP), and Ilan Koren (Weizmann Institute of Science, Israel).

The findings show that freezing, which in this case occurs not at 0 °C, as it does in our day-to-day lives, but at about -10 °C, depends mainly on a combination of three factors: atmospheric humidification, solar radiation, and aerosols. In southern Amazonia’s rainy season (roughly December-April), the atmosphere is extremely clear and the origin of the particles in the aerosols is natural. They come from the condensation of gases emitted by the forest, and from wind abrasion of soil and vegetation. They typically contain pollen, microorganisms, and sea salt, among other kinds of particles. In the burning season, which occurs annually in August-October, large-scale fires emit a huge amount of smoke, which spreads throughout the region and is blown by the wind to other regions. “They produce much worse pollution than urban activities in the city of São Paulo, for example,” Correia said.

The study contributes to the knowledge of the behavior of clouds in the Amazon and can be enriched by further research. “The influence of clouds on the climate is very important. This is the most complex topic in climate models that set out to forecast what will happen with regard to this theme in the future, so any improvement in knowledge of how clouds function is a major contribution to the advancement of climate science,” he stressed.

Today, machine learning, artificial intelligence, and algorithmic advancements made by research scientists and engineers are driving more targeted medical therapies through the power of prediction. The ability to rapidly analyze large amounts of complex data has clinicians closer to providing individualized treatments for patients, intending to create better outcomes through more proactive, personalized medicine and care.  

“In medicine, we need to be able to make predictions,” said John F. McDonald, professor in the School of Biological Sciences and director of the Integrated Cancer Research Center in the Petit Institute for Bioengineering and Bioscience at the Georgia Institute of Technology. One way is through understanding cause and reflecting relationships, as a cancer patient’s response to drugs, he explained. The other way is through correlation. 

“In analyzing complex datasets in cancer biology, we can use machine learning, which is simply a sophisticated way to look for correlations. The advantage is that computers can look for these correlations in extremely large and complex data sets.”

Now, McDonald’s team and the Ovarian Cancer Institute are using ensemble-based machine learning algorithms to predict how patients will respond to cancer-fighting drugs with high accuracy rates. The results of their most recent work have been published in the Journal of Oncology Research.  

For the study, McDonald and his colleagues developed predictive machine learning-based models for 15 distinct cancer types, using data from 499 independent cell lines provided by the National Cancer Institute. Those models were then validated against a clinical dataset containing seven chemotherapeutic drugs, administered either singularly or in combination, to 23 ovarian cancer patients. The researchers found an overall predictive accuracy of 91%.

“While additional validation will need to be carried out using larger numbers of patients with multiple types of cancer,” McDonald noted, “our preliminary finding of 90% accuracy in the prediction of drug responses in ovarian cancer patients is extremely promising and gives me hope that the days of being able to accurately predict optimal cancer drug therapies for individual patients is in sight."

The study was conducted in collaboration with the Ovarian Cancer Institute (OCI) in Atlanta, where McDonald serves as a chief research officer. Other authors are Benedict Benigno, MD (OCI founder and chief executive officer, as well as an obstetrician-gynecologist, surgeon, and oncologist); Nick Housley, a postdoctoral researcher in McDonald’s Georgia Tech lab; and the paper’s lead author, Jai Lanka, an intern with OCI. 

The challenges in predicting cancer treatments

The complex nature of cancer makes it a challenging problem when it comes to predicting drug responses, McDonald said. Patients with the same type of cancer will often respond differently to the same treatment. 

“Part of the problem is that the cancer cell is a highly integrated network of pathways and patient tumors that display the same characteristics clinically may be quite different on the molecular level,” he explained. 

A major goal of personalized cancer medicine is to accurately predict likely responses to drug treatments based upon genomic profiles of individual patient tumors. 

“In our approach, we utilize an ensemble of machine learning methods to build predictive algorithms — based on correlations between gene expression profiles of cancer cell lines or patient tumors with previously observed responses — to a variety of cancer drugs. The future goal is that gene expression profiles of tumor biopsies can be fed into the algorithms, and likely patient responses to different drug therapies can be predicted with high accuracy,” said McDonald.   

Machine learning is already being applied to the data coming from the genomic profiles of tumor biopsies, but before the researchers’ work, these methods have typically involved a single algorithmic approach. 

McDonald and his team decided to combine several algorithm approaches that use multiple ways to analyze complex data; one even uses a three-dimensional approach. They found using this ensemble-based approach significantly boosted predictive accuracy.

The algorithms the team used have names like Support Vector Machines (SVM), Random Forest classifier (RF), K-Nearest Neighbor classifier (KNN), and Logistic Regression classifier (LR). 

“They’re all fairly technical, and they’re all different computational mathematical approaches, and all of them are looking for correlations,” said McDonald. “It’s just a question of which one to use, and for different data sets, we find that one model might work better than another.”

However, more patient datasets that combine genomic profiles with responses to cancer drugs are needed to advance the research.  

“If we want to have a clinical impact, we must validate our models using data from a large number of patients,” said McDonald, who added that many datasets are held by pharmaceutical companies who use them in drug development. That data is typically considered proprietary, private information. And although a significant amount of genomic data of cancer patients is generally available, it’s not typically correlated with patient responses to drugs.

McDonald is currently talking with medical insurance companies about access to relevant datasets, as well. “It costs insurance companies a significant amount of money to pay for drug treatments that don’t work,” he noted. Time, medical fees, and ultimately, many lives could be saved by providing researchers with these types of information. 

“Right now, a percentage of patients will not respond to a drug, but we don’t know that until after six weeks of chemotherapy,” said McDonald. “What we hope is that we will soon have tools that can accurately predict the probability of a patient responding to first-line therapies — and if they don’t respond, to be able to make accurate predictions as to the next drug to be tried.”