TLCrys: Transfer Learning Based Method for Protein Crystallization Prediction

X-ray diffraction technique is one of the most common methods of ascertaining protein structures, yet only 2–10% of proteins can produce diffraction-quality crystals. Several computational methods have been proposed so far to predict protein crystallization. Nevertheless, the current state-of-the-art computational methods are limited by the scarcity of experimental data. Thus, the prediction accuracy of existing models hasn’t reached the ideal level. To address the problems above, we propose a novel transfer-learning-based framework for protein crystallization prediction, named TLCrys. The framework proceeds in two steps: pre-training and fine-tuning. The pre-training step adopts attention mechanism to extract both global and local information of the protein sequences. The representation learned from the pre-training step is regarded as knowledge to be transferred and fine-tuned to enhance the performance of crystalization prediction. During pre-training, TLCrys adopts a multi-task learning method, which not only improves the learning ability of protein encoding, but also enhances the robustness and generalization of protein representation. The multi-head self-attention layer guarantees that different levels of the protein representation can be extracted by the fine-tuned step. During transfer learning, the fine-tuning strategy used by TLCrys improves the task-specialized learning ability of the network. Our method outperforms all previous predictors significantly in five crystallization stages of prediction. Furthermore, the proposed methodology can be well generalized to other protein sequence classification tasks.

The dual function of impurity on protein crystallization

It is widely believed that high purity protein is critical to crystallization, and crystallization is facilitated when protein is pure enough. Impurity can induce lattice strain, internal stress, and rotation...

The Role of Cations of the Precipitant in the Interaction of Protein Molecules in the Lysozyme Oligomers in Crystallization Solutions

At the moment, the main opinion is that protein crystallization depends mainly on the the precipitant anions, therefore, there have been only few works devoted to the problem of the influence of its cations. Using the molecular dynamics method, we investigated the stability, changes in the compactness and structural transformations of lysozyme dimers and octamers in solutions with different precipitants (LiCl, NaCl, KCl and CuCl2) in order to study the contribution of cations during crystal formation in more detail. As a result, we found that cations have a rather noticeable effect on the behavior of oligomers: the higher the atomic mass of the cation, the greater the changes in the dimers structures during its dynamics and, according to the data of SAXS experiments, the lower the concentration of dimers. However, for octamers, this dependence is more complicated.

Ionic Liquids as Protein Crystallization Additives

Among its attributes, the mythical philosopher’s stone is supposedly capable of turning base metals to gold or silver. In an analogous fashion, we are finding that protein crystallization optimization using ionic liquids (ILs) often results in the conversion of base protein precipitate to crystals. Recombinant inorganic pyrophosphatases (8 of the 11 proteins) from pathogenic bacteria as well as several other proteins were tested for optimization by 23 ILs, plus a dH2O control, at IL concentrations of 0.1, 0.2, and 0.4 M. The ILs were used as additives, and all proteins were crystallized in the presence of at least one IL. For 9 of the 11 proteins, precipitation conditions were converted to crystals with at least one IL. The ILs could be ranked in order of effectiveness, and it was found that ~83% of the precipitation-derived crystallization conditions could be obtained with a suite of just eight ILs, with the top two ILs accounting for ~50% of the hits. Structural trends were found in the effectiveness of the ILs, with shorter-alkyl-chain ILs being more effective. The two top ILs, accounting for ~50% of the unique crystallization results, were choline dihydrogen phosphate and 1-butyl-3-methylimidazolium tetrafluoroborate. Curiously, however, a butyl group was present on the cation of four of the top eight ILs.

Engineering an efficient and bright split Corynactis californica green fluorescent protein

Split green fluorescent protein (GFP) has been used in a panoply of cellular biology applications to study protein translocation, monitor protein solubility and aggregation, detect protein–protein interactions, enhance protein crystallization, and even map neuron contacts. Recent work shows the utility of split fluorescent proteins for large scale labeling of proteins in cells using CRISPR, but sets of efficient split fluorescent proteins that do not cross-react are needed for multiplexing experiments. We present a new monomeric split green fluorescent protein (ccGFP) engineered from a tetrameric GFP found in Corynactis californica, a bright red colonial anthozoan similar to sea anemones and scleractinian stony corals. Split ccGFP from C. californica complements up to threefold faster compared to the original Aequorea victoria split GFP and enable multiplexed labeling with existing A. victoria split YFP and CFP.