Artificial Intelligence Medical Compendium

Motivation: Peptide toxicity remains a major bottleneck in drug development, as small sequence variations can drastically alter cytotoxic profiles. Experimental assessment is costly and time-consuming, motivating computational predictors. Existing ap... read more 

Transposable elements (TEs) constitute a substantial fraction of the human genome and contribute to genomic and epigenomic regulation. However, TE-related knowledge remains fragmented across the literature and databases, with inconsistent identifiers... read more 

Predictive grid cells, a recently discovered functional class of neurons in medial entorhinal cortex (MEC), exhibit grid-like tuning, not with respect to where an animal currently is, but where it will be. Such coding, which is disjoint from standard... read more 

Cancers frequently co-opt genes and transcriptional networks associated with embryos and stem cells, yet the extent and significance of expressing gene programs associated with the earliest stages of embryonic development in cancer remain unknown. Sy... read more 

Background: Marker-based estimates of completeness and contamination are widely used to identify near-complete metagenome-assembled genomes (MAGs), but they do not detect foreign sequences joined within a contig. Results: Across 63 samples from eight... read more 

Models trained on protein sequences can predict mutational effects, but their outputs represent a convolution of different phenotypes. Isolating stability-independent activity effects is essential for rational enzyme design, allosteric drug design an... read more 

In pathology, pattern recognition relies on a frame of reference built through extensive training and experience, which may still fall short when encountering rare cancer subtypes. Archives of previously diagnosed tumors often contain relevant preced... read more 

Whole-slide images may contain coexisting lesions, yet weak supervision typically provides only the most severe diagnosis. We formulate this setting as dominant-label multiple instance learning, where the observed label is the maximum-severity elemen... read more 

Contrastive vision-language pre-training supports zero-shot chest X-ray diagnosis, but standard objectives treat every off-diagonal image-report pair in a batch as a negative. In radiology, recurring findings across patients mean that instance mismat... read more 

Genetic and drug perturbation screens are important tools for elucidating disease mechanisms and validating therapeutic targets, but experimental costs limit the number of candidates that can be tested. Existing perturbation screening methods typical... read more