Dataset Information

Cluster learning-assisted directed evolution.

ABSTRACT: Directed evolution, a strategy for protein engineering, optimizes protein properties (i.e., fitness) by expensive and time-consuming screening or selection of large mutational sequence space. Machine learning-assisted directed evolution (MLDE), which screens sequence properties in silico, can accelerate the optimization and reduce the experimental burden. This work introduces a MLDE framework, cluster learning-assisted directed evolution (CLADE), that combines hierarchical unsupervised clustering sampling and supervised learning to guide protein engineering. The clustering sampling selectively picks and screens variants in targeted subspaces, which guides the subsequent generation of diverse training sets. In the last stage, accurate predictions via supervised learning models improve final outcomes. By sequentially screening 480 sequences out of 160,000 in a four-site combinatorial library with five equal experimental batches, CLADE achieves the global maximal fitness hit rate up to 91.0% and 34.0% for GB1 and PhoQ datasets, respectively, improved from 18.6% and 7.2% obtained by random-sampling-based MLDE.

SUBMITTER: Qiu Y

PROVIDER: S-EPMC9267417 | biostudies-literature | 2021 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Cluster learning-assisted directed evolution.

Qiu Yuchi Y Hu Jian J Wei Guo-Wei GW

Nature computational science 20211209 12

Directed evolution, a strategy for protein engineering, optimizes protein properties (i.e., fitness) by expensive and time-consuming screening or selection of large mutational sequence space. Machine learning-assisted directed evolution (MLDE), which screens sequence properties <i>in silico</i>, can accelerate the optimization and reduce the experimental burden. This work introduces a MLDE framework, cluster learning-assisted directed evolution (CLADE), that combines hierarchical unsupervised cl ...[more]

PMID: 35811998

Similar Datasets

Project description:The presence of defects, such as pores, in materials processed using additive manufacturing represents a challenge during the manufacturing of many engineering components. Recently, ultrasonic vibration-assisted (UV-A) directed energy deposition (DED) has been shown to reduce porosity, promote grain refinement, and enhance mechanical performance in metal components. Whereas it is evident that the formation of such microstructural features is affected by the melt pool behavior, the specific mechanisms by which ultrasonic vibration (UV) influences the melt pool remain elusive. In the present investigation, UV was applied in situ to DED of 316L stainless steel single tracks and bulk parts. For the first time, high-speed video imaging and thermal imaging were implemented in situ to quantitatively correlate the application of UV to melt pool evolution in DED. Extensive imaging data were coupled with in-depth microstructural characterization to develop a statistically robust dataset describing the observed phenomena. Our findings show that UV increases the melt pool peak temperature and dimensions, while improving the wettability of injected particles with the melt pool surface and reducing particle residence time. Near the substrate, we observe that UV results in a 92% decrease in porosity, and a 54% decrease in dendritic arm spacing. The effect of UV on the melt pool is caused by the combined mechanisms of acoustic cavitation, ultrasound absorption, and acoustic streaming. Through in situ imaging we demonstrate quantitatively that these phenomena, acting simultaneously, effectively diminish with increasing build height and size due to acoustic attenuation, consequently decreasing the positive effect of implementing UV-A DED. Thus, this research provides valuable insight into the value of in situ imaging, as well as the effects of UV on DED melt pool dynamics, the stochastic interactions between the melt pool and incoming powder particles, and the limitations of build geometry on the UV-A DED technique.

Dataset Information

Cluster learning-assisted directed evolution.

Publications

Cluster learning-assisted directed evolution.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets