Sex and gender differences and biases in artificial intelligence for biomedicine and healthcare.

Precision Medicine implies a deep understanding of inter-individual differences in health and disease that are due to genetic and environmental factors. To acquire such understanding there is a need for the implementation of different types of technologies based on artificial intelligence (AI) that enable the identification of biomedically relevant patterns, facilitating progress towards individually tailored preventative and therapeutic interventions. Despite the significant scientific advances achieved so far, most of the currently used biomedical AI technologies do not account for bias detection. Furthermore, the design of the majority of algorithms ignore the sex and gender dimension and its contribution to health and disease differences among individuals. Failure in accounting for these differences will generate sub-optimal results and produce mistakes as well as discriminatory outcomes. In this review we examine the current sex and gender gaps in a subset of biomedical technologies used in relation to Precision Medicine. In addition, we provide recommendations to optimize their utilization to improve the global health and disease landscape and decrease inequalities.

Evidence for a conserved inhibitory binding mode between the membrane fusion assembly factors Munc18 and syntaxin in animals.

The membrane fusion necessary for vesicle trafficking is driven by the assembly of heterologous SNARE proteins orchestrated by the binding of Sec1/Munc18 (SM) proteins to specific syntaxin SNARE proteins. However, the precise mode of interaction between SM proteins and SNAREs is debated, as contrasting binding modes have been found for different members of the SM protein family, including the three vertebrate Munc18 isoforms. While different binding modes could be necessary, given their roles in different secretory processes in different tissues, the structural similarity of the three isoforms makes this divergence perplexing. Although the neuronal isoform Munc18a is well-established to bind tightly to both the closed conformation and the N-peptide of syntaxin 1a, thereby inhibiting SNARE complex formation, Munc18b and -c, which have a more widespread distribution, are reported to mainly interact with the N-peptide of their partnering syntaxins and are thought to instead promote SNARE complex formation. We have reinvestigated the interaction between Munc18c and syntaxin 4 (Syx4). Using isothermal titration calorimetry, we found that Munc18c, like Munc18a, binds to both the closed conformation and the N-peptide of Syx4. Furthermore, using a novel kinetic approach, we found that Munc18c, like Munc18a, slows down SNARE complex formation through high-affinity binding to syntaxin. This strongly suggests that secretory Munc18s in general control the accessibility of the bound syntaxin, probably preparing it for SNARE complex assembly.

DNA free energy-based promoter prediction and comparative analysis of Arabidopsis and rice genomes.

The cis-regulatory regions on DNA serve as binding sites for proteins such as transcription factors and RNA polymerase. The combinatorial interaction of these proteins plays a crucial role in transcription initiation, which is an important point of control in the regulation of gene expression. We present here an analysis of the performance of an in silico method for predicting cis-regulatory regions in the plant genomes of Arabidopsis (Arabidopsis thaliana) and rice (Oryza sativa) on the basis of free energy of DNA melting. For protein-coding genes, we achieve recall and precision of 96% and 42% for Arabidopsis and 97% and 31% for rice, respectively. For noncoding RNA genes, the program gives recall and precision of 94% and 75% for Arabidopsis and 95% and 90% for rice, respectively. Moreover, 96% of the false-positive predictions were located in noncoding regions of primary transcripts, out of which 20% were found in the first intron alone, indicating possible regulatory roles. The predictions for orthologous genes from the two genomes showed a good correlation with respect to prediction scores and promoter organization. Comparison of our results with an existing program for promoter prediction in plant genomes indicates that our method shows improved prediction capability.