No abstract

SummaryIn this study, the eff ects of temperature-shift (from 30 to 25 °C) and temperature-constant (at 30 °C) cultivation on the mass of Monascus fuliginosus CG-6 mycelia and concentration of the produced monacolin K (MK) were monitored. The expression levels of the MK biosynthetic genes of M. fuliginosus CG-6 at constant and variable culture temperatures were analysed by real-time quantitative polymerase chain reaction (RT-qPCR). The total protein was collected and determined by liquid chromatography-electrospray ionisation with tandem mass spectrometry (LC-ESI-MS/MS). Results showed that the maximum mycelial mass in temperature-shift cultivation was only 0.477 g of dry cell mass per dish, which was lower than that in temperature-constant cultivation (0.581 g of dry cell mass per dish); however, the maximum concentration of MK in temperature-shift cultivation (34.5 μg/mL) was 16 times higher than that in temperature-constant cultivation at 30 °C (2.11 μg/mL). Gene expression analysis showed that the expression of the MK biosynthetic gene cluster at culture temperature of 25 °C was higher than that at 30 °C, which was similar to the trend of the MK concentration, except for individual MK B and MK C genes. Analysis of diff erential protein exp ression revealed that 2016 proteins were detected by LC-ESI-MS/MS. The expression level of effl ux pump protein coded by the MK I gene exhibited the same upregulated trend as the expression of MK I in temperature-shift cultivation. Temperature-shift cultivation enhanced the expression of proteins in the secondary metabolite production pathway, but suppressed the expression of proteins involved in the mycelial growth.

Deep stacked RNNs are usually hard to train. Adding shortcut connections across different layers is a common way to ease the training of stacked networks. However, extra shortcuts make the recurrent step more complicated. To simply the stacked architecture, we propose a framework called shortcut block, which is a marriage of the gating mechanism and shortcuts, while discarding the selfconnected part in LSTM cell. We present extensive empirical experiments showing that this design makes training easy and improves generalization. We propose various shortcut block topologies and compositions to explore its effectiveness. Based on this architecture, we obtain a 6% relatively improvement over the state-of-the-art on CCGbank supertagging dataset. We also get comparable results on POS tagging task.

Combinatory Category Grammar (CCG) supertagging is a task to assign lexical categories to each word in a sentence. Almost all previous methods use fixed context window sizes to encode input tokens. However, it is obvious that different tags usually rely on different context window sizes. This motivates us to build a supertagger with a dynamic window approach, which can be treated as an attention mechanism on the local contexts. We find that applying dropout on the dynamic filters is superior to the regular dropout on word embeddings. We use this approach to demonstrate the state-of-the-art CCG supertagging performance on the standard test set.