New machine-learning method reveals sex-specific genes and pathways in Alzheimer’s disease

Researchers at Baylor College of Medicine and the Jan and Dan Duncan Neurological Research Institute have made a significant breakthrough in understanding Alzheimer’s disease (AD). They have developed a novel machine-learning method called Evolutionary Action Machine Learning (EAML) that enables the identification of sex-specific genes and molecular pathways associated with AD. The study, published in Nature Communications, reveals that females experience faster cognitive decline and cerebral atrophy, while males have higher mortality rates in relation to AD.

Dr. Olivier Lichtarge, one of the lead researchers, explains that the EAML software utilizes an advanced computational predictive metric called the evolutionary action (EA) score to identify genetic factors that influence AD risk separately in males and females. By leveraging evolutionary data efficiently, the researchers were able to analyze smaller cohorts and accurately identify genes involved in the sex-specific differences observed in AD.

The EAML approach encompasses nine machine learning algorithms and focuses on analyzing the functional impact of non-synonymous coding variants, which are DNA mutations that affect the resulting protein’s structure and function. These variants are evaluated using the EA score to estimate their deleterious effect on biological processes.

The team employed EAML to study coding variants in 2,729 AD patients and 2,441 control subjects, leading to the identification of 98 genes associated with AD. Among these genes were known players in AD biology, reinforcing the value of combining machine learning with evolutionary information to uncover genes and pathways relevant to complex diseases like AD. Furthermore, the team found that these genes exhibited abnormal expression in AD brains and formed functional connections. The pathways influenced by these genes were associated with neuroinflammation, as well as microglial and astrocytic biology, aligning with their potential involvement in AD pathophysiology.

To validate their findings, the researchers collaborated with other experts in the field and utilized fruit fly models of AD. Through advanced behavioral testing methods, they identified 36 genes that modulated tau-induced degeneration and 29 genes that modulated Aβ42-induced neurodegeneration. Impressively, nine of these genes were able to ameliorate the neurodegeneration caused by both Tau and Aβ42, two proteins known to accumulate in AD patients. This not only confirmed the functional relevance of the identified genes in neurodegeneration but also suggested potential therapeutic targets for AD treatment.

In summary, the researchers at Baylor College of Medicine and the Duncan NRI have developed EAML, a machine-learning method that effectively identifies sex-specific genes and pathways involved in AD. This approach, combined with in vivo studies using fruit fly models, has provided valuable insights into the mechanisms of AD and highlighted potential therapeutic avenues for further exploration.

In their study, the researchers aimed to investigate the sex-specific differences in Alzheimer’s disease (AD) by utilizing the EAML analysis separately for males and females within the cohort. The results revealed 157 AD-associated genes in males and 127 in females. Notably, the genes identified through sex-separated analysis showed stronger connections to known AD GWAS genes compared to those identified through combined sex studies. This suggests that conducting sex-separated analysis enhances the sensitivity in identifying AD-associated genes and improves the ability to predict disease risk.

Furthermore, the researchers discovered that certain biological pathways might have a more pronounced impact on AD development in one sex compared to the other. For instance, female-specific EAML candidates were found to be associated with a module related to cell cycle control and DNA quality control. This led to the identification of a group of neuroprotective genes in females linked to the BRCA1 gene, known for its association with breast cancer. These findings indicate potential biological connections between AD and breast cancer, two diseases more prevalent in females. The implications of these findings extend to the development of therapeutic strategies and the design of sex-stratified clinical trials for AD.

Additionally, the EAML method demonstrated consistent and robust predictive capabilities even with smaller sample sizes. Even when tested with only 700 samples, EAML was able to recover over 50% of the candidates identified in the entire dataset, outperforming currently used predictive algorithms. This improved predictive capability of EAML enables researchers to achieve accurate and reliable predictions with smaller datasets. This breakthrough paves the way for incorporating sex-specific analyses into disease-gene association studies that may not have yielded reliable results using conventional methods.

The researchers believe that the success of EAML in identifying new targets for AD not only provides fresh insights into the genetic factors influencing the disorder but also emphasizes the importance of systematically applying sex-specific analyses in disease-gene association studies. This innovative approach has the potential to revolutionize our understanding of complex diseases like AD and drive the development of personalized treatments tailored to individuals’ genetic makeup.

The study involved several other researchers from institutions including Baylor College of Medicine, Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, and UTHealth McGovern Medical School.

Source: Texas Children’s Hospital

Leave a Comment