bioRxiv Science⌕ Search

Biology subjects

Krawczuk, I.

Publications and source records attributed to Krawczuk, I..

2 recordsLinked to original sources
bioRxiv2025

Deep-learning-assisted SICM for enhanced real-time imaging of nanoscale biological dynamics

Scanning Ion Conductance Microscopy (SICM) provides high-resolution, nanoscale imaging of living cells, but it is generally limited by a slow scan rate, making it challenging to capture dynamic processes in real time. To tackle this challenge, we propose an integrated data acquisition and computational framework that improves the temporal resolution of SICM by selectively skipping certain scan lines. A partial convolutional neural network (Partial-CNN) model is developed and trained on SICM images and their corresponding masks to reconstruct the complete images from the under-sampled data, ensuring the retention of structural integrity. This approach significantly reduces the image acquisition time (i.e., by 30-60%) without compromising quality, as validated through multiple quantitative metrics. Compared to conventional deep learning methods, the Partial-CNN demonstrates higher accuracy in reconstructing fine details and maintaining consistent height maps across skipped regions. We show that this method provides an increased temporal resolution and retains image fidelity, making it suitable for real-time dynamic SICM imaging and improving the smart scanning microscopy applications in time-resolved biological imaging. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=92 SRC="FIGDIR/small/663539v1_ufig1.gif" ALT="Figure 1"> View larger version (37K): org.highwire.dtl.DTLVardef@562d32org.highwire.dtl.DTLVardef@1b12dcaorg.highwire.dtl.DTLVardef@13535f3org.highwire.dtl.DTLVardef@6904f0_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioengineering↗
bioRxiv2025

Crowdsourced Protein Design: Lessons From the Adaptyv EGFR Binder Competition

In this report, we summarize and analyze the 2024 Adaptyv protein design competition. Participants used computational and Machine Learning (ML) methods of their choice to design proteins that bind the Epidermal Growth Factor Receptor (EGFR), a key drug target involved in cell growth, differentiation, and cancer development. Over 1,800 designs were submitted across two rounds. Of these, 601 proteins were selected and characterized for expression and binding affinity to EGFR, with competitors both optimizing existing binders (KD = 1.21 nM) and creating de novo binders (KD = 82 nM). All selected designs were experimentally validated using Adaptyvs automated Bio-Layer Interferometry (BLI) pipeline. This competition illustrates the potential of crowdsourcing to drive creativity and innovation in protein design. However, it also exposed key challenges, such as the lack of standardized benchmarks, experimental design targets, and robust computational metrics for method comparison. We anticipate that future competitions will address these gaps and further motivate progress in computational protein design.

bioengineering↗