The amount of available biological data, originating from various experimental procedures (genome and transcriptome sequencing, structure determination, functional assays), is vast. In our group, we employ computational techniques, such as deep learning, molecular dynamics simulations, and sequence analysis to make use of this data. We are particularly interested in understanding how protein folds have emerged and how protein structures and functions are encoded by the alphabet of 20 amino acids. For details see projects tab.
Oct. 14, 2021
The evolutionary origin of slipknot topology in proteins remains a mystery. In a collaborative effort with Joanna Sułkowska’s group [https://jsulkowska.cent.uw.edu.pl/ ], we identified a slipknot motif in a family of two-domain membrane transporters, which enabled us to direct investigation of its evolution. Moreover, we developed tools that may be useful in the investigation of the evolution of other repeated-domain proteins. The work was published by PLoS Computational Biology.
Sept. 24, 2021
The Rossmann fold enzymes are involved in many essential biochemical pathways such as nucleotide and amino acid metabolism. Their activity is strictly dependent on the presence of a nucleotide-based cofactor bound to a conserved region called beta-alpha-beta core. By applying deep-learning techniques we created a computational tool that enables the prediction and re-design of cofactor specificity in Rossmann fold enzymes. Our work was appreciated by prestigious Briefings in Bioinformatics. The method is available as a web server and as a standalone tool .
May 7, 2021
Rossmann fold is a large class of proteins encompassing a multitude of various oxidoreductases and methyltransferases. All of them employ a conserved region, termed the core motif, to interact with nucleotide-based cofactors (NAD, NADP, FAD, and SAM) that are essential for their functioning. In course of our research, we developed a computational tool for the prediction and redesign of such protein-cofactor interactions. The tool can be accessed via the web server and GitHub repository. A detailed description of the methodology can be found in a preprint that just showed up online.
Feb. 3, 2021
KfrA genes are widely-present in broad-host-range plasmids and have been shown to act as transcriptional auto-regulators. We have recently taken part in a study that focused on the structure and function of coiled-coil domains of two KfrA proteins. Using tools developed in our lab, SamCC Turbo and DeepCoil we revealed peculiar structural features of these domains that may be essential for the functioning of KfrA proteins. Further molecular dynamics and modeling simulations have shown that these specific non-canonical regions are constantly present, suggesting that they are an inherent feature of KfrA coiled-coil domains. Overall, performed studies allowed us to postulate the hypothesis that KfrA type proteins have moonlighting activity – they not only act as transcriptional auto-regulators but form cytoskeletal structures, which might facilitate plasmid DNA delivery and its position in the cells.
Jan. 11, 2021
Coiled coils are widespread, highly-regular, protein domains involved in diverse processes ranging from providing structural rigidity to the transduction of conformational changes. We have developed SamCC Turbo, a fully automated tool for the detection and characterization of these biologically important structures. By applying SamCC Turbo to the whole Protein Data Bank served via localpdb, we built a comprehensive library of coiled-coil domains. We showed how this machine learning-ready data set can be used in focused and global analyses, such as the development of a new version of DeepCoil, a tool for the detection of coiled-coil domains in sequences.
Oct. 26, 2020
We are delighted to inform you that two members of our team Adriana Bukała and Kamil Kamiński have defended their theses. Adriana's BSc work entitled "Ancestral state reconstruction approach in investigating Rossmann fold evolution" focused on deciphering the evolution of regions responsible for the cofactor binding, whereas Kamil's MSc work entitled "Re-engineering cofactor specificity of the Rossmann fold proteins using graph neural networks" describes a new method for predicting the cofactor specificity in Rossmann enzymes. Congratulations!
April 14, 2020
The disulfide bond formation catalyzed by bacterial proteins of the Dsb (disulfide bond) system is a crucial step in the folding process of a protein. Although we now have a good understanding of the Escherichia coli disulfide bond formation system, there are significant gaps in our knowledge concerning the Dsb systems of other bacteria, including Campylobacter jejuni, a food-borne, zoonotic pathogen. In a recent article prof. Katarzyna Jagusztyn-Krynicka and her co-workers investigated C8J_1298 protein, a Dsb system component from this organism. Our group contributed to this study by performing a phylogenetic analysis of C8J_1298 and its homologs and thus shedding light on its evolution.
Dec. 3, 2019
We are happy to inform that our colleague, Jan Ludwiczak, has obtained an award for PhD students founded by the President of Polish Academy of Sciences. More details (in Polish) can be found here. Congratulations!
Aug. 28, 2019
The last ~30 amino acids of kinesin-14 Ncd head play a crucial but still poorly understood role in a lever-like rotation of a long superhelical 'stalk' segment while one of the kinesin's heads is bound to the microtubule. The members of our lab Jan Ludwiczak and Antonio Marinho published a paper describing how a new conformational state can be accessed and stabilized with several hydrogen bonds mediated by residues located on the border or within the C-terminal linker. Such stable compact conformation of Ncd may represent an initial state of the working stroke.
Aug. 2, 2019
It is our pleasure to inform that Aleksander Wiński was awarded Diamond Grant that will allow him to start his PhD studies in the fall. Moreover, Jan Ludwiczak obtained Etiuda grant during which he will visit our collaborators in Max Planck Institute for Developmental Biology in Tübingen.
July 5, 2019
Jun 10-14 we were honored to host two young students of Polish Children's Fund. During the stay, they had an opportunity to participate in ever-day lab life and work on their own project. The students, Emilia and Krzysztof, have developed a deep-learning based predictor of protein secondary structure, which can be accessed online. In the last day, we got a beautiful hand-made postcard. Thus, we believe that the students enjoyed the stay as much as we enjoyed hosting them :-).
May 6, 2019
As you may have heard, protein structures consist of secondary structure elements. The most abundant are alpha-helices and beta-sheets; however, there are more types of secondary structure elements, for example, pi-helices. Pi-helices are similar to alpha-helices but are less stable. In our recent work we showed that these rare secondary structure elements are frequently associated with functionalities such as ligand- or ion-binding and, most importantly, we presented a tool - PiPred - that can predict pi-helices directly from a sequence. PiPred is available as a web service and can be downloaded as a standalone tool.
Dec. 29, 2018
One of the main challenges of protein flexibility analyses is the mathematical representation of the protein conformational space. Our team member Antonio Marinho recently published a paper describing a new protein dynamic analysis method based on differential geometry representation of protein backbone. Presented method avoids most pitfalls of popular mathematical representation, such as phi-psi, atomic coordinates and collective variables. The FleXgeo software to compute such analyses is available for download.
Dec. 29, 2018
RNA-recognition motif (RRM) is an RNA-interacting protein domain that plays an important role in the processes of RNA metabolism such as splicing, editing, export, degradation, and regulation of translation. We developed RRMdb database, which allows for a quick identification and annotation of RRM domains in a given protein sequence. We provide the RRMdb database as a tool both for the experimentalists searching for basic knowledge about RNA-binding proteins with RRM domains as well as those studying the evolution of RRM domains. RRMdb can be accessed here and the paper is available here.
Dec. 29, 2018
That's one small step for man, one giant leap for coiled coils! By combining our expertise on coiled-coil domains with the modern deep learning approaches, we have built DeepCoil, a fast and sensitive method for the detection of coiled coils based on sequence data. DeepCoil can be accessed via MPI Toolkit or downloaded as a stand-alone software. We plan to further develop DeepCoil, thus all suggestions on how to improve it will be more than welcome. The paper is available here.
Sept. 3, 2018
We are happy to announce that Aleksander Wiński obtained BSc degree from the Faculty of Physics, University of Warsaw. The title of his work was "The use of deep learning methods to predict protein π-helix secondary structure based on sequence and evolutionary information" (Polish title: "Zastosowanie metod uczenia głębokiego do przewidywania białkowych struktur drugorzędowych typu helisa π na podstawie sekwencji i informacji ewolucyjnej"). The work (only in Polish) can be downloaded here.
Aug. 8, 2018
We are very happy to announce that Foundation for Polish Science awarded Stanisław Dunin-Horkawicz with First Team grant for the project entitled "An evolutionary-based approach to design the substrate specificity of the Rossmann fold enzymes".
July 3, 2018
Sequence design combined with machine learning techniques comprise a powerful tool for aiding sequence-structure relationship studies. In our work "Variability of the core geometry in parallel coiled-coil bundles" recently published in Journal of Structural Biology, we applied this approach to study determinants of hydrophobic core packing geometry in parallel homotetrameric coiled coils.
June 1, 2018
π-helices are short, unstable secondary structure elements present in 25% of all protein structures, often in functionally important regions. Given their similarity to α-helices, prediction of π-helices is a challenging task and none of the currently available secondary structure prediction methods tackle it. To fill this void, we have developed PiPred, a publicly available tool for predicting π-helices in protein sequences. PiPred is freely available on the web (as a part of the Quick2D tool). A standalone version is available for download at GitHub.
April 21, 2018
The collaboration with Prof. Katarzyna Jagusztyn-Krynicka from Faculty of Biology (University of Warsaw) resulted in a new publication that just has appeared in PLoS ONE (link). Our phylogenetic analysis helped to put the experimental work of Prof. Jagusztyn-Krynicka and her co-workers into an evolutionary context.
Feb. 14, 2018
We are happy to announce that our paper "Combining Rosetta with molecular dynamics (MD): a benchmark of the MD-based ensemble protein design" has been published in Journal of Structural Biology. The work has been done in collaboration with Dr. Adam Jarmuła from Nencki Institute of Experimental Biology. The paper is available online and more information can be found on the project's website at GitHub.
Jan. 3, 2018
CCpred is a server for prediction of helix axial rotation from sequence. You can acess it from Software tab on our webpage. Documentation will be available soon.
Nov. 28, 2017
There are still many sections that demand more content but we will be filling them soon. Fresh news about our group will be placed in Home section.
Jan Ludwiczak, PhD, PhD Student
Favorite research & tools: My scientific interests revolve around proteins, their sequences, structures and dynamics. Favorite techniques I use are machine learning, molecular dynamics and protein design.
Alma mater: Łódź University of Technology
Other interests: brewing, computer games and music
Jędrzej Kubica, BSc, MSc Student
Favorite research & tools: I study chemistry and bioinformatics. I love taking part in hackathons.
Alma mater: University of Warsaw, Faculty of Chemistry
Other interests: cars, languages, windsurfing
Kamil Krakowski, MSc, PhD Student
Favorite research & tools: The primary area of my scientific interests is molecular biology, with particular emphasis on RNA and transcriptional modulation of gene expression. I am also passionate about microbiology and bioinformatics.
Alma mater: University of Warsaw, Faculty of Biology
Other interests: Programming, watching movies, biking, travelling
Aleksander Wiński, BSc, MSc Student
Favorite research & tools: I am interested in bioinformatics, especially in application of machine learning to protein sequence and structure analysis
Alma mater: University of Warsaw, Faculty of Physics
Other interests: Movies, table tennis, chess
Stanisław Dunin-Horkawicz, PhD, Lab Leader
Favorite research & tools: My main research interest is the application of computational tools to understand biological complexity. Our team focuses on deciphering the sequence-structure-function relationships of proteins.
Alma mater: Max Planck Institute for Developmental Biology / The University of Tübingen
Other interests: Biking, books and cats.
Kamil Kamiński, MSc, MSc Student
Favorite research & tools: Data Science applications to biological and medical data, programming and statistical modeling
Alma mater: Faculty of Physics, University of Warsaw
Other interests: historical books, brazilian ju jitsu and music
Rafał Madaj, PhD, Postdoc
Favorite research & tools: I am focused on computer-aided drug design, especially employing molecular docking and molecular dynamics simulations.
Alma mater: Lodz University of Technology, Faculty of Biotechnology and Food Sciences
Other interests: Brewing, chess, weightlifting