< /i>In silico< //i> prediction of < /i>in vitro< //i> protein liquid-liquid phase separation experiments outcomes with multi-head neural attention

Raimondi, Daniele; Orlando, Gabriele; Michiels, Emiel; Pakravan, Donya; Bratek-Skicki, Anna; Van den Bosch, Ludo; Moreau, Yves; Rousseau, Frederic; Schymkowitz, Joost

doi:10.1093/bioinformatics/btab350

< /i>In silico< //i> prediction of < /i>in vitro< //i> protein liquid-liquid phase separation experiments outcomes with multi-head neural attention

Author:

Raimondi, Daniele

Orlando, Gabriele ; Michiels, Emiel ; Pakravan, Donya ; Bratek-Skicki, Anna ; Van den Bosch, Ludo ; Moreau, Yves ; Rousseau, Frederic ; Schymkowitz, Joost

Keywords:

Science & Technology, Life Sciences & Biomedicine, Technology, Physical Sciences, Biochemical Research Methods, Biotechnology & Applied Microbiology, Computer Science, Interdisciplinary Applications, Mathematical & Computational Biology, Statistics & Probability, Biochemistry & Molecular Biology, Computer Science, Mathematics, STADIUS-21-91, I011220N#55634482, I011620N#55793044, G0H1716N#53816276, 01 Mathematical Sciences, 06 Biological Sciences, 08 Information and Computing Sciences, Bioinformatics, 31 Biological sciences, 46 Information and computing sciences, 49 Mathematical sciences

Abstract:

MOTIVATION: Proteins able to undergo liquid-liquid phase separation (LLPS) in vivo and in vitro are drawing a lot of interest, due to their functional relevance for cell life. Nevertheless, the proteome-scale experimental screening of these proteins seems unfeasible, because besides being expensive and time-consuming, LLPS is heavily influenced by multiple environmental conditions such as concentration, pH and temperature, thus requiring a combinatorial number of experiments for each protein. RESULTS: To overcome this problem, we propose a neural network model able to predict the LLPS behavior of proteins given specified experimental conditions, effectively predicting the outcome of in vitro experiments. Our model can be used to rapidly screen proteins and experimental conditions searching for LLPS, thus reducing the search space that needs to be covered experimentally. We experimentally validate Droppler's prediction on the TAR DNA-binding protein in different experimental conditions, showing the consistency of its predictions. AVAILABILITY AND IMPLEMENTATION: A python implementation of Droppler is available at https://bitbucket.org/grogdrinker/droppler. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.