Lightweight deep neural network for radio frequency interference detection and segmentation in synthetic aperture radar.

Radio frequency interference (RFI) poses challenges in the analysis of synthetic aperture radar (SAR) images. Existing RFI suppression systems rely on prior knowledge of the presence of RFI. This paper proposes a lightweight neural network-based algorithm for detecting and segmenting RFI (LDNet) in the time-frequency domain. The network accurately delineates RFI pixel regions in time-frequency spectrograms. To mitigate the impact on the operational speed of the entire RFI suppression system, lightweight modules and pruning operations are introduced. Compared to threshold-based RFI detection algorithms, deep learning-based segmentation networks, and AC-UNet specifically designed for RFI detection, LDNet achieves improvements in mean intersection over union (MIoU) by 24.56%, 13.29%, and 7.54%, respectively.Furthermore, LDNet reduces model size by 99.03% and inference latency by 24.53% compared to AC-UNet.

Thiazolidine Deprotection Using an Organic Solvent Extractable Aldehyde Scavenger for One-Pot Four-Segment Ligation.

We report a simple and convenient N-terminal thiazolidine (Thz) deprotection strategy and its application in one-pot multisegment ligation. In this strategy, O-benzylhydroxylamine (O-BHA) is used to efficiently and rapidly convert Thz into N-terminal cysteine. O-BHA can be easily separated from the ligation buffer by organic solvent extraction, avoiding the degradation of the peptide thioester by O-BHA. The utility of the O-BHA-based one-pot ligation strategy has been demonstrated in the assembly of CC chemokine ligand-2.

Thioproline-Based Oxidation Strategy for Direct Preparation of N-Terminal Thiazolidine-Containing Peptide Thioesters from Peptide Hydrazides.

We describe a simple and robust oxidation strategy for preparing N-terminal thiazolidine-containing peptide thioesters from peptide hydrazides. We find for the first time that l-thioproline can be used as a protective agent to prevent the nitrosation of N-terminal thiazolidine during peptide hydrazide oxidation. The thioproline-based oxidation strategy has been successfully applied to the chemical synthesis of CC chemokine ligand-2 (69aa) and omniligase-C (113aa), thereby demonstrating its utility in hydrazide-based native chemical ligation.

Employing unnatural promiscuity of sortase to construct peptide macrocycle libraries for ligand discovery.

With the increasing attention paid to macrocyclic scaffolds in peptide drug development, genetically encoded peptide macrocycle libraries have become invaluable sources for the discovery of high-affinity peptide ligands targeting disease-associated proteins. The traditional phage display technique of constructing disulfide-tethered macrocycles by cysteine oxidation has the inherent drawback of reduction instability of the disulfide bond. Chemical macrocyclization solves the problem of disulfide bond instability, but the involved highly electrophilic reagents are usually toxic to phages and may bring undesirable side reactions. Here, we report a unique Sortase-mediated Peptide Ligation and One-pot Cyclization strategy (SPLOC) to generate peptide macrocycle libraries, avoiding the undesired reactions of electrophiles with phages. The key to this platform is to mine the unnatural promiscuity of sortase on the X residue of the pentapeptide recognition sequence (LPXTG). Low reactive electrophiles are incorporated into the X-residue side chain, enabling intramolecular cyclization with the cysteine residue of the phage-displayed peptide library. Utilizing the genetically encoded peptide macrocycle library constructed by the SPLOC platform, we found a high-affinity bicyclic peptide binding TEAD4 with a nanomolar KD value (63.9 nM). Importantly, the binding affinity of the bicyclic peptide ligand is 102-fold lower than that of the acyclic analogue. To our knowledge, this is the first time to mine the unnatural promiscuity of ligases to generate peptide macrocycles, providing a new avenue for the construction of genetically encoded cyclic peptide libraries.