I have confirmation from several sources now that all T32s, many F30s and F31s, and most or all Center awards (P30, P50) have been terminated at Columbia.

This is quite damaging to research and to individuals.

This is pure terrorism and cannot be legal. But litigation will take time...

Finally, we discuss downstream applications of models to understand disease-relevant non-coding variants, such as functionally informed fine-mapping and de novo variant prioritization. 4/4

We also review variant effect prediction evaluations that have been performed to date on genomic deep learning models, highlighting strengths and limitations of current models and the need for more comprehensive evaluation. 3/4

We cover two popular genomic deep learning modeling paradigms — supervised sequence-to-activity models and self-supervised genomic language models — and describe practical considerations for using both types of models to make variant effect predictions. 2/4