Search for: All records

Abstract Adding synthetic nucleotides to DNA increases the linear information density of DNA molecules. Here we report that it also can increase the diversity of their three-dimensional folds. Specifically, an additional nucleotide (dZ, with a 5-nitro-6-aminopyridone nucleobase), placed at twelve sites in a 23-nucleotides-long DNA strand, creates a fairly stable unimolecular structure (that is, the folded Z-motif, or fZ-motif) that melts at 66.5 °C at pH 8.5. Spectroscopic, gel and two-dimensional NMR analyses show that the folded Z-motif is held together by six reverse skinny dZ⁻:dZ base pairs, analogous to the crystal structure of the free heterocycle. Fluorescence tagging shows that the dZ⁻:dZ pairs join parallel strands in a four-stranded compact down–up–down–up fold. These have two possible structures: one with intercalated dZ⁻:dZ base pairs, the second without intercalation. The intercalated structure would resemble the i-motif formed by dC:dC⁺-reversed pairing at pH ≤ 6.5. This fZ-motif may therefore help DNA form compact structures needed for binding and catalysis. 

ABSTRACT The interplay between defense and counterdefense systems of bacteria and bacteriophages has been driving the evolution of both organisms, leading to their great genetic diversity. Restriction-modification systems are well-studied defense mechanisms of bacteria, while phages have evolved covalent modifications as a counterdefense mechanism to protect their genomes against restriction. Here, we present evidence that these genome modifications might also have been selected to counter, broadly, the CRISPR-Cas systems, an adaptive bacterial defense mechanism. We found that the phage T4 genome modified by cytosine hydroxymethylation and glucosylation (ghmC) exhibits various degrees of resistance to the type V CRISPR-Cas12a system, producing orders of magnitude more progeny than the T4(C) mutant, which contains unmodified cytosines. Furthermore, the progeny accumulated CRISPR escape mutations, allowing rapid evolution of mutant phages under CRISPR pressure. A synergistic effect on phage restriction was observed when two CRISPR-Cas12a complexes were targeted to independent sites on the phage genome, another potential countermechanism by bacteria to more effectively defend themselves against modified phages. These studies suggest that the defense-counterdefense mechanisms exhibited by bacteria and phages, while affording protection against one another, also provide evolutionary benefits for both. IMPORTANCE Restriction-modification (R-M) and CRISPR-Cas systems are two well-known defense mechanisms of bacteria. Both recognize and cleave phage DNA at specific sites while protecting their own genomes. It is well accepted that T4 and other phages have evolved counterdefense mechanisms to protect their genomes from R-M cleavage by covalent modifications, such as the hydroxymethylation and glucosylation of cytosine. However, it is unclear whether such genome modifications also provide broad protection against the CRISPR-Cas systems. Our results suggest that genome modifications indeed afford resistance against CRISPR systems. However, the resistance is not complete, and it is also variable, allowing rapid evolution of mutant phages that escape CRISPR pressure. Bacteria in turn could target more than one site on the phage genome to more effectively restrict the infection of ghmC-modified phage. Such defense-counterdefense strategies seem to confer survival advantages to both the organisms, one of the possible reasons for their great diversity. 

Conversational search is one of the ultimate goals of information retrieval. Recent research approaches conversational search by simplified settings of response ranking and conversational question answering, where an answer is either selected from a given candidate set or extracted from a given passage. These simplifications neglect the fundamental role of retrieval in conversational search. To address this limitation, we introduce an open-retrieval conversational question answering (ORConvQA) setting, where we learn to retrieve evidence from a large collection before extracting answers, as a further step towards building functional conversational search systems. We create a dataset, OR-QuAC, to facilitate research on ORConvQA. We build an end-to-end system for ORConvQA, featuring a retriever, a reranker, and a reader that are all based on Transformers. Our extensive experiments on OR-QuAC demonstrate that a learnable retriever is crucial for ORConvQA. We further show that our system can make a substantial improvement when we enable history modeling in all system components. Moreover, we show that the reranker component contributes to the model performance by providing a regularization effect. Finally, further in-depth analyses are performed to provide new insights into ORConvQA. 

Conversational AI is a rapidly developing research field in both industry and academia. As one of the major branches of conversational AI, question answering and conversational search has attracted significant attention of researchers in the information retrieval community. It has been a long overdue feature for search engines or conversational assistants to retrieve information iteratively and interactively in a conversational manner. Previous work argues that conversational question answering (ConvQA) is a simplified but concrete setting of conversational search. In this setting, one of the major challenges is to leverage the conversation history to understand and answer the current question. In this work, we propose a novel solution for ConvQA that involves three aspects. First, we propose a positional history answer embedding method to encode conversation history with position information using BERT (Bidirectional Encoder Representations from Transformers) in a natural way. BERT is a powerful technique for text representation. Second, we design a history attention mechanism (HAM) to conduct a "soft selection" for conversation histories. This method attends to history turns with different weights based on how helpful they are on answering the current question. Third, in addition to handling conversation history, we take advantage of multi-task learning (MTL) to do answer prediction along with another essential conversation task (dialog act prediction) using a uniform model architecture. MTL is able to learn more expressive and generic representations to improve the performance of ConvQA. We demonstrate the effectiveness of our model with extensive experimental evaluations on QuAC, a large-scale ConvQA dataset. We show that position information plays an important role in conversation history modeling. We also visualize the history attention and provide new insights into conversation history understanding. The complete implementation of our model will be open-sourced.