Rational Protein Engineering Guided by Deep Mutational Scanning

Shin, HyeonSeok,Cho, Byung-Kwan MDPI 2015 INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES Vol.16 No.9

<P>Sequence–function relationship in a protein is commonly determined by the three-dimensional protein structure followed by various biochemical experiments. However, with the explosive increase in the number of genome sequences, facilitated by recent advances in sequencing technology, the gap between protein sequences available and three-dimensional structures is rapidly widening. A recently developed method termed deep mutational scanning explores the functional phenotype of thousands of mutants via massive sequencing. Coupled with a highly efficient screening system, this approach assesses the phenotypic changes made by the substitution of each amino acid sequence that constitutes a protein. Such an informational resource provides the functional role of each amino acid sequence, thereby providing sufficient rationale for selecting target residues for protein engineering. Here, we discuss the current applications of deep mutational scanning and consider experimental design.</P>

Double-attention mechanism of sequence-to-sequence deep neural networks for automatic speech recognition

육동석,임단,유인철 한국음향학회 2020 韓國音響學會誌 Vol.39 No.5

Sequence-to-sequence deep neural networks with attention mechanisms have shown superior performance across various domains, where the sizes of the input and the output sequences may differ. However, if the input sequences are much longer than the output sequences, and the characteristic of the input sequence changes within a single output token, the conventional attention mechanisms are inappropriate, because only a single context vector is used for each output token. In this paper, we propose a double-attention mechanism to handle this problem by using two context vectors that cover the left and the right parts of the input focus separately. The effectiveness of the proposed method is evaluated using speech recognition experiments on the TIMIT corpus.

Sequence to Sequence based LSTM (LSTM-s2s)모형을 이용한 댐유입량 예측에 대한 연구

한희찬,최창현,정재원,김형수 한국수자원학회 2021 한국수자원학회논문집 Vol.54 No.3

Forecasting dam inflow based on high reliability is required for efficient dam operation. In this study, deep learning technique, which is one of the data-driven methods and has been used in many fields of research, was manipulated to predict the dam inflow. The Long Short-Term Memory deep learning with Sequence-to-Sequence model (LSTM-s2s), which provides high performance in predicting time-series data, was applied for forecasting inflow of Soyang River dam. Various statistical metrics or evaluation indicators, including correlation coefficient (CC), Nash-Sutcliffe efficiency coefficient (NSE), percent bias (PBIAS), and error in peak value (PE), were used to evaluate the predictive performance of the model. The result of this study presented that the LSTM-s2s model showed high accuracy in the prediction of dam inflow and also provided good performance for runoff event based runoff prediction. It was found that the deep learning based approach could be used for efficient dam operation for water resource management during wet and dry seasons. 효율적인 댐 운영을 위해서는 높은 신뢰도를 기반으로 하는 유입량 예측이 요구된다. 본 연구에서는 최근 다양한 분야에서 사용되고 있는 데이터 기반의 예측 방법 중 하나인 딥러닝을 댐 유입량 예측에 활용하였다. 그 중 시계열 자료 예측에 높은 성능을 보이는 Sequence-to-Sequence 구조 기반의 Long Short-Term Memory 딥러닝 모형(LSTM-s2s)을 이용하여 소양강 댐의 유입량을 예측하였다. 모형의 예측 성능을 평가하기 위해 상관계수, Nash–Sutcliffe 효율계수, 평균편차비율, 그리고 첨두값 오차를 이용하였다. 그 결과, LSTM-s2s 모형은 댐 유입량 예측에 대한 높은 정확도를 보였으며, 단일 유량 수문곡선 기반의 예측 성능에서도 높은 신뢰도를 보였다. 이를 통해 홍수기와 이수기에 수자원 관리를 위한 효율적인 댐 운영에 딥러닝 모형의 적용 가능성을 확인할 수 있었다.

Compound <i>EGFR</i> mutation is frequently detected with co-mutations of actionable genes and associated with poor clinical outcome in lung adenocarcinoma

Kim, Eun Young,Cho, Eun Na,Park, Heae Surng,Hong, Ji Young,Lim, Seri,Youn, Jong Pil,Hwang, Seung Yong,Chang, Yoon Soo LANDES BIOSCIENCE 2016 Cancer Biology & Therapy Vol. No.

<P>Compound <I>EGFR</I> mutations, defined as double or multiple mutations in the <I>EGFR</I> tyrosine kinase domain, are frequently detected with advances in sequencing technology but its clinical significance is unclear. This study analyzed 61 cases of <I>EGFR</I> mutation positive lung adenocarcinoma using next-generation sequencing (NGS) based repeated deep sequencing panel of 16 genes that contain actionable mutations and investigated clinical implication of compound <I>EGFR</I> mutations. Compound <I>EGFR</I> mutation was detected in 15 (24.6%) of 61 cases of <I>EGFR</I> mutation-positive lung adenocarcinoma. The majority (12/15) of compound mutations are combination of the atypical mutation and typical mutations such as exon19 deletion, L858R or G719X substitutions, or exon 20 insertion whereas 3 were combinations of rare atypical mutations. The patients with compound mutation showed shorter overall survival than those with simple mutations (83.7 <I>vs</I>. 72.8 mo; <I>P</I> = 0.020, Breslow test). Among the 115 missense mutations discovered in the tested genes, a few number of actionable mutations were detected irrelevant to the subtype of <I>EGFR</I> mutations, including <I>ALK rearrangement</I>, <I>BCL2L11</I> intron 2 deletion, <I>KRAS</I> c.35G>A<I>, PIK3CA</I> c.1633G>A which are possible target of crizotinib, BH3 mimetics, <I>MEK</I> inhibitors, and <I>PI3K-tyrosine kinase inhibitors</I>, respectively. 31 missense mutations were detected in the cases with simple mutations whereas 84 in those with compound mutation, showing that the cases with compound missense mutation have higher burden of missense mutations (<I>P</I> = 0.001, independent sample <I>t</I>-test). Compound <I>EGFR</I> mutations are detected at a high frequency using NGS-based repeated deep sequencing. Because patients with compound <I>EGFR</I> mutations showed poor clinical outcomes, they should be closely monitored during follow-up.</P>

Linear-Time Korean Morphological Analysis Using an Action-based Local Monotonic Attention Mechanism

Hwang, Hyunsun,Lee, Changki Electronics and Telecommunications Research Instit 2020 ETRI Journal Vol.42 No.1

For Korean language processing, morphological analysis is a critical component that requires extensive work. This morphological analysis can be conducted in an end-to-end manner without requiring a complicated feature design using a sequence-to-sequence model. However, the sequence-to-sequence model has a time complexity of O(n²) for an input length n when using the attention mechanism technique for high performance. In this study, we propose a linear-time Korean morphological analysis model using a local monotonic attention mechanism relying on monotonic alignment, which is a characteristic of Korean morphological analysis. The proposed model indicates an extreme improvement in a single threaded environment and a high morphometric F1-measure even for a hard attention model with the elimination of the attention mechanism formula.

miRNA Pattern Discovery from Sequence Alignment

Xiaohan Sun,Junying Zhang 한국정보처리학회 2017 Journal of information processing systems Vol.13 No.6

MiRNA is a biological short sequence, which plays a crucial role in almost all important biological process. MiRNA patterns are common sequence segments of multiple mature miRNA sequences, and they are ofsignificance in identifying miRNAs due to the functional implication in miRNA patterns. In the proposedapproach, the primary miRNA patterns are produced from sequence alignment, and they are then cut intoshort segment miRNA patterns. From the segment miRNA patterns, the candidate miRNA patterns areselected based on estimated probability, and from which, the potential miRNA patterns are further selectedaccording to the classification performance between authentic and artificial miRNA sequences. Threeparameters are suggested that bi-nucleotides are employed to compute the estimated probability of segmentmiRNA patterns, and top 1% segment miRNA patterns of length four in the order of estimated probabilitiesare selected as potential miRNA patterns.

Identification of miR172 family members and their putative targets responding to drought stress in Solanum tuberosum

Eul-Won Hwang,신선주,박수철,Mi-Jeong Jeong,권혁빈 한국유전학회 2011 Genes & Genomics Vol.33 No.2

MicroRNAs (miRNAs) function as negative posttranscriptional regulators during plant development and in response to biotic and abiotic stress. In this study, we identified drought stress related miRNAs from sequencing reads obtained from second-generation sequencing. This method is useful for determination of miRNA expression profiles when genomic sequences are not available and for identification of putative miRNA by aligning reads with sequences from miRBase. Here, we present identification of a family of drought responsive miRNAs from potato plants, stu-miR172c,stu-miR172d, and stu-miR172e, and show expression profiles obtained in response to drought stress as well as their target mRNAs.

miRNA Pattern Discovery from Sequence Alignment

Sun, Xiaohan,Zhang, Junying Korea Information Processing Society 2017 Journal of information processing systems Vol.13 No.6

MiRNA is a biological short sequence, which plays a crucial role in almost all important biological process. MiRNA patterns are common sequence segments of multiple mature miRNA sequences, and they are of significance in identifying miRNAs due to the functional implication in miRNA patterns. In the proposed approach, the primary miRNA patterns are produced from sequence alignment, and they are then cut into short segment miRNA patterns. From the segment miRNA patterns, the candidate miRNA patterns are selected based on estimated probability, and from which, the potential miRNA patterns are further selected according to the classification performance between authentic and artificial miRNA sequences. Three parameters are suggested that bi-nucleotides are employed to compute the estimated probability of segment miRNA patterns, and top 1% segment miRNA patterns of length four in the order of estimated probabilities are selected as potential miRNA patterns.

miR171 Family Members are Involved in Drought Response in Solanum tuberosum

Hwang, Eul-Won,Shin, Seon-Ju,Yu, Byeong-Kuk,Byun, Myung-Ok,Kwon, Hawk-Bin 한국식물학회 2011 Journal of Plant Biology Vol.54 No.1

MicroRNAs (miRNAs) are a group of small noncoding RNAs found in both animals and plants. miRNAs function as negative post-transcriptional regulators during plant development and in response to biotic and abiotic stress. In this work, we identified drought stress-related miRNAs from reads obtained from second-generation sequencing. This method is useful to determine miRNA expression profiles when genomic sequences are not available and to find putative miRNAs by aligning reads with sequences from miRBase. Here, we report the identification of a family of drought-responsive miRNAs, stu-miR171a, stu-miR171b, and stu-miR171c, frompotato plants, their expression profiles upon drought stress, and their target mRNAs.

심전도 기반 사용자 인증 시스템의 인증 속도 및 효율성 개선을 위한 시퀀스 투 시퀀스 구조의 비전 트랜스포머

김상규,유선국,강희철 한국멀티미디어학회 2024 멀티미디어학회논문지 Vol.27 No.6

This study aims to improve the authentication speed and efficiency of user authentication based on Electrocardiogram (ECG) signals. ECG signals are unique to each individual and are difficult to replicate and manipulate, making them a promising candidate for new authentication methods. However, compared to fingerprints, iris, and facial recognition, ECG-based authentication has the inconvenience of requiring longer measurement times for acquiring the necessary ECG signals. To address these challenges, we developed a novel approach that integrates a sequence-to-sequence Transformer architecture with the Vision Transformer (ViT) model to achieve enhanced authentication speed and efficiency. The Sequence-to-Sequence structure allows input ECG signals to be processed in short patches by the transformer model, enabling real-time analysis of electrocardiogram signals. In addition, we employed masking techniques to tokenized input patches, enabling the model to authenticate early even when only partial ECG signals are available. The model’s performance at a sequence length of 3 is as follows: accuracy 98.8%, precision 100%, recall 97.62%, and F1-score 98.8%. We expect that by inputting three bits of the measured electrocardiogram signals into the proposed model, it will be able to authenticate users quickly and accurately.