Clinical feature-related single-base substitution sequence signatures identified with an unsupervised machine learning approach

Background Mutation processes leave different signatures in genes. For single-base substitutions, previous studies have suggested that mutation signatures are not only reflected in mutation bases but also in neighboring bases. However, because of the lack of a method to identify features of long sequences next to mutation bases, the understanding of how flanking sequences influence mutation signatures is limited. Methods We constructed a long short-term memory-self organizing map (LSTM-SOM) unsupervised neural network. By extracting mutated sequence features via LSTM and clustering similar features with the SOM, single-base substitutions in The Cancer Genome Atlas database were clustered according to both their mutation site and flanking sequences. The relationship between mutation sequence signatures and clinical features was then analyzed. Finally, we clustered patients into different classes according to the composition of the mutation sequence signatures by the K-means method and then studied the differences in clinical features and survival between classes. Results Ten classes of mutant sequence signatures (mutation blots, MBs) were obtained from 2,141,527 single-base substitutions via LSTM-SOM machine learning approach. Different features in mutation bases and flanking sequences were revealed among MBs. MBs reflect both the site and pathological features of cancers. MBs were related to clinical features, including age, sex, and cancer stage. The class of an MB in a given gene was associated with survival. Finally, patients were clustered into 7 classes according to the MB composition. Significant differences in survival and clinical features were observed among different patient classes. Conclusions We provided a method for analyzing the characteristics of mutant sequences. Result of this study showed that flanking sequences, together with mutation bases, shape the signatures of SBSs. MBs were shown related to clinical features and survival of cancer patients. Composition of MBs is a feasible predictive factor of clinical prognosis. Further study of the mechanism of MBs related to cancer characteristics is suggested.

Clinical Feature-Related Single-Base Substitution Sequence Signatures Identified with an Unsupervised Machine Learning Approach

Background: Mutation processes leave different signatures in genes. For single-base substitutions, previous studies have suggested that mutation signatures are not only reflected in mutation bases but also in neighboring bases. However, because of the lack of a method to identify features of long sequences next to mutation bases, the understanding of how flanking sequences influence mutation signatures is limited.Methods: We constructed a long short-term memory – self organizing map (LSTM-SOM) unsupervised neural network. By extracting mutated sequence features via LSTM and clustering similar features with the SOM, single-base substitutions in The Cancer Genome Atlas database were clustered according to both their mutation site and flanking sequences. The relationship between mutation sequence signatures and clinical features was then analyzed. Finally, we clustered patients into different classes according to the composition of the mutation sequence signatures by the K-means method and then studied the differences in clinical features and survival between classes.Results: Ten classes of mutant sequence signatures (mutation blots, MBs) were obtained from 2,141,527 single-base substitutions via LSTM-SOM machine learning approach. Different features in mutation bases and flanking sequences were revealed among MBs. MBs reflect both the site and pathological features of cancers. MBs were related to clinical features, including age, gender, and cancer stage. The class of an MB in a given gene was associated with survival. Finally, patients were clustered into 7 classes according to the MB composition. Significant differences in survival and clinical features were observed among different patient classes.Conclusions: We provided a method for analyzing the characteristics of mutant sequences. Result of this study showed that flanking sequences, together with mutation bases, shape the signatures of SBSs. MBs were shown related to clinical features and survival of cancer patients. Composition of MBs is a feasible predictive factor of clinical prognosis. Further study of the mechanism of MBs related to cancer characteristics is suggested.

In Vivo Modeling of Clonal Competition Using CRISPR-Based Gene Editing Reveals Novel Fitness Variables in Multiple Myeloma

Introduction: Multiple Myeloma (MM) is an incurable malignancy characterized by the proliferation of clonal plasma cells in the bone marrow (BM). MM almost always progresses from the precursor states of monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM), which indicates the presence of a gradual clonal evolution underlying progression from the original stages of tumor development to the time of clinical presentation. Clonal heterogeneity adds another layer of complexity to that, by introducing interclonal competition in the context of disease progression or therapeutic bottlenecks. Here we developed a mouse model to investigate the impact of multiple clonal mutations on tumor development, as well as the competitive expansion of individual clones. Methods: Primary mouse MM Vk*Myc cells stably expressing Cas9 were infected with validated sgRNAs to knockout (KO) genes of interest (P53, Cyld, Rb1, Dis3, Prdm1, Traf3 and Fam46c) that are significantly mutated in human MM. KO cells were mixed at a 1:1 ratio with control cells infected with control sgRNA and injected intravenously into 8-week-old RAG2 KO mice. Vk*Myc cells were then isolated from bone marrow and spleen through CD138 positive selection, followed by genomic DNA extraction and NGS sequencing to understand the dynamic changes in abundance of mutants from injection to early and late timepoints. Results: In vitro, most knockout Vk*Myc cells had a similar proliferation rate to control cells with the exception of P53 and Rb1 knockout cells, which grew faster as expected; both P53 and RB1 are known cell cycle regulators. However, when co-injected into RAG2 KO mice (Vk*Myc cells constructed with Cas9 do not engraft in C57BL/6 mice), although P53 and Rb1 knockout cells remained the strongest competitors, occupying the majority of the tumor, most KO cells exhibited significantly enhanced proliferation over control cells. These results indicate that certain mutations only become advantageous in the context of the tumor microenvironment, while mutations that directly affect the tumor cell's proliferation rate give rise to more flexible, potent clones. To better understand these differences, we took advantage of the CRISPR-induced heterogeneous pool of genomic edits per gene, and looked at clonal abundancy rates within each knockout population separately. Interestingly, we found mutants with certain insertions/deletions grew faster than others and were overrepresented at the late stage of disease, even when they were generated from the same double-stranded break. Although it is well established that mutations in different regions of the same gene might have different effects, these results indicate that different mutations in the exact same spot can give rise to clones of variable potency and beg the question of whether mutation sequence is as important as mutation hotspot. Conclusion: We established a mouse model to study clonal competition in vivo, utilizing the CRISPR-Cas9 genome editing toolset. Through our model, we were able to witness a range of competitive potential among genes that are significantly mutated in multiple myeloma, with P53- and RB1-mutants as the strongest competitors. Furthermore, we observed that competitive potential can be conditional, with certain mutants conferring fitness advantage only in the context of tumor microenvironment. Adding another layer of complexity to differential fitness, we found that different mutations in the same spot of the same gene give rise to clones of varied potency, implicating mutation sequence as a novel fitness variable. In this study, we thus demonstrate that mutational candidates can be prioritized based on competitive potential, a process of the utmost importance given multiple myeloma's marked genetic heterogeneity. Disclosures Ghobrial: Celgene: Consultancy; Takeda: Consultancy; Janssen: Consultancy; BMS: Consultancy.

The Diagnostic Method of Concrete Dam Monitoring Information Valid Interval

On the basis of the analysis of the displacement of concrete dam and its related influential factors, based on the evolvement of nonlinear dynamics of concrete dam, it can effectively identify the mutations position of measured value and the attribute interval of dynamical system applied with the wavelet analysis, dynamic structural mutation theory and other numerical analysis methods. When detecting after separating structural mutation sequence, it can finally get the relative stable displacement time series of dynamical structure, so it can realize the diagnostic separation of the monitoring information effective interval. At the end of the paper, through applying a certain concrete arch dam, it is proved that the proposed method of concrete dam mutations diagnosis of is of great significance for the real-time monitoring of the workability state of a dam.