Jisheng Liang

Mr. Jisheng Liang

Shanghai Jiao Tong University, Shanghai, China
orcid

Research Focus & Metrics

Academic Career Length

22Years (2005-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
WCWenxiang Cai
Wenxiang Cai
WSWeichen Song
Weichen Song
ZLZhe Liu
Zhe Liu
DMDhruba Tara Maharjan
Dhruba Tara Maharjan
JLJisheng Liang
Jisheng Liang
NGNing Lin Guan
Ning Lin Guan
Schizophrenia (SCZ) is a severe mental disorder that may result in hallucinations, delusions, and extremely disordered thinking. How each cell type in the brain contributes to SCZ occurrence is still unclear. Here, we leveraged the human dorsolateral prefrontal cortex bulk RNA-seq data, then used the RNA-seq deconvolution algorithm CIBERSORTx to generate SCZ brain single-cell RNA-seq data for a comprehensive analysis to understand SCZ-associated brain cell types and gene expression changes. Firstly, we observed that the proportions of brain cell types in SCZ differed from normal samples. Among these cell types, astrocyte, pericyte, and PAX6 cells were found to have a higher proportion in SCZ patients (astrocyte: SCZ = 0.163, control = 0.145, P.adj = 4.9 × 10−4, effect size = 0.478; pericyte: SCZ = 0.057, control = 0.066, P.adj = 1.1 × 10−4, effect size = 0.519; PAX6: SCZ = 0.014, control = 0.011, P.adj = 0.014, effect size = 0.377), while the L5/6_IT_CAR3 cells and LAMP5 cells are the exact opposite (L5/6_IT_Car3: SCZ = 0.102, control = 0.108, P.adj = 0.016, effect size = 0.369; LAMP5: SCZ = 0.057, control = 0.066, P.adj = 2.2 × 10−6, effect size = 0.617). Next, we investigated gene expression in cell types and functional pathways in SCZ. We observed chemical synaptic transmission dysregulation in two types of GABAergic neurons (PVALB and LAMP5), and immune reaction involvement in GABAergic neurons (SST) and non-neuronal cell types (endothelial and oligodendrocyte). Furthermore, we observed that some differential expression genes from bulk RNA-seq displayed cell-type-specific abnormalities in the expression of molecules in SCZ. Finally, the cell types with the SCZ-related transcriptomic changes could be considered to belong to the same module since we observed two major similar coordinated transcriptomic changes across these cell types. Together, our results offer novel insights into cellular heterogeneity and the molecular mechanisms underlying SCZ.

3 October 2022 in Biology
Verified
Journal Article
Open access
PDF file
ZLZhe Liu
Zhe Liu
WPWeihao Pan
Weihao Pan
WLWeihao Li
Weihao Li
XZXuyang Zhen
Xuyang Zhen
JLJisheng Liang
Jisheng Liang
NGNing Lin Guan
Ning Lin Guan
Though AlphaFold2 has attained considerably high precision on protein structure prediction, it is reported that directly inputting coordinates into deep learning networks cannot achieve desirable results on downstream tasks. Thus, how to process and encode the predicted results into effective forms that deep learning models can understand to improve the performance of downstream tasks is worth exploring. In this study, we tested the effects of five processing strategies of coordinates on two single-sequence protein binding site prediction tasks. These five strategies are spatial filtering, the singular value decomposition of a distance map, calculating the secondary structure feature, and the relative accessible surface area feature of proteins. The computational experiment results showed that all strategies were suitable and effective methods to encode structural information for deep learning models. In addition, by performing a case study of a mutated protein, we showed that the spatial filtering strategy could introduce structural changes into HHblits profiles and deep learning networks when protein mutation happens. In sum, this work provides new insight into the downstream tasks of protein-molecule interaction prediction, such as predicting the binding residues of proteins and estimating the effects of mutations.

1 June 2005 in ACM SIGKDD Explorations Newsletter
Verified
Journal Article
Open access
JLJisheng Liang
Jisheng Liang
KKKrzysztof Koperski
Krzysztof Koperski
TNThien Nguyen
Thien Nguyen
GMGiovanni Marchisio
Giovanni Marchisio
We present a framework that bridges the gap between natural language processing (NLP) and text mining. Central to this is a new approach to text parameterization that captures many interesting attributes of text usually ignored by standard indices, like the term-document matrix. By storing NLP tags, the new index supports a higher degree of knowledge discovery and pattern finding from text. The index is relatively compact, enabling dynamic search of arbitrary relationships and events in large document collections. We can export search results in formats and data structures that are transparent to statistical analysis tools like S-PLUSID®. In a number of experiments, we demonstrate how this framework can turn mountains of unstructured information into informative statistical graphs.

18 August 2022 in Preprints.org
Verified
Preprint
Open access
PDF file
WCWenxiang Cai
Wenxiang Cai
WSWeichen Song
Weichen Song
ZLZhe Liu
Zhe Liu
DMDhruba Tara Maharjan
Dhruba Tara Maharjan
JLJisheng Liang
Jisheng Liang
NGNing Lin Guan
Ning Lin Guan
Schizophrenia (SCZ) is a severe mental disorder that may result in hallucinations, delusions, and extremely disordered thinking. How each cell type in the brain contributes to SCZ occurrence is still unclear. Here, we leveraged the human dorsolateral prefrontal cortex bulk RNA-seq data, then used the RNA-seq deconvolution algorithm CIBERSORTx to generate SCZ brain single-cell RNA-seq data for a comprehensive analysis to understand SCZ-associated brain cell types and gene expression changes. Firstly, we observed that the proportions of brain cell types in SCZ differed from normal samples. Among these cell types, astrocyte, pericyte, and PAX6 cells were found to have a higher proportion in SCZ patients (astrocyte: SCZ = 0.163, Control = 0.145, P.adj = 4.9×10-4; pericyte: SCZ = 0.057, Control = 0.066, P.adj = 1.1×10-4; PAX6 : SCZ = 0.014, Control = 0.011, P.adj = 0.014), while the L5/6_IT_CAR3 cells and LAMP5 cells are the exact opposite (L5/6_IT_Car3 : SCZ = 0.102, Control = 0.108, P.adj = 0.016; LAMP5 : SCZ = 0.057, Control = 0.066, P.adj = 2.2×10-6). Next, we investigated gene expression in cell types and functional pathways in SCZ. We observed chemical synaptic transmission dysregulation in two types of GABAergic neurons (PVALB and LAMP5), and immune reaction involvement in GABAergic neurons (SST) and non-neuronal cell types (endothelial and oligodendrocyte). Furthermore, we observed that some differential expression genes from bulk RNA-seq displayed cell-type-specific abnormal in the expression of molecules in SCZ. Finally, the cell types with the SCZ-related transcriptomic changes could be considered to belong to the same module since we observed two major similar coordinated transcriptomic changes across these cell types. Together, our results offer novel insights into cellular heterogeneity and the molecular mechanisms underlying SCZ.

20 October 2023
Verified
Conference Paper
Open access
JLJisheng Liang
Jisheng Liang
ZLZhe Liu
Zhe Liu
WCWenxiang Cai
Wenxiang Cai
YBYihang Bao
Yihang Bao
NGNing Lin Guan
Ning Lin Guan
Protein-nucleic acid interactions, pivotal to the pathogenesis of certain diseases, have been the focus of intensive study over the years. Protein missense mutations can impact protein-nucleic acid binding affinity. Existing machine learning tools for predicting these effects rely heavily on protein structural information, limiting their applicability when only protein sequences are available. In this study, a deep learning-based model, Transformers-RNP, was designed to work solely on sequential information and mutation data. To enhance performance, we integrated HHblits-predicted evolutionarily conserved properties into the model. We trained Transformers-RNP on a recently published database for protein-nucleic acid binding affinities, ProNAB. Our model obtained a Root Mean Square Error (RMSE) of 0.809 Kcal/mol and a Pearson's correlation coefficient of 0.728 on the test dataset. Furthermore, a case study was performed on a protein-nucleic acid complex with single-point mutation. Notably, our model's weights were focused on the interaction region. The results show that Transformers-RNP serves as a robust tool for discerning the mutation effects on protein-RNA interaction sites.