Michał Startek
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Michał Startek has a total of 112 co-authors. Here are the top co-authors by number of shared publications:
- Anna Gambin
Institute of Informatics, University of Warsaw, Stefana Banacha 2, Warsaw 02-097, Poland
15 shared publications - Błażej Miasojedow
University of Warsaw
5 shared publications - Mateusz Krzysztof Łącki
University of Warsaw
5 shared publications - Dariusz Grzebelus
Department of Plant Biology and Biotechnology, Faculty of Biotechnology and Horticulture, University of Agriculture in Krakow, Krakow, Poland
4 shared publications Dirk Valkenborg
Interuniversity Institute of Biostatistics and Statistical Bioinformatics, Hasselt University, BE3500 Hasselt, Belgium
4 shared publicationsEwa P. Gajewska
3 shared publications
Latest Publications
- 22 October 2026
- Published by The R Foundation
opentimsr: An Open-Source Loader for Bruker's timsTOF Data Files
Reads raw data from Bruker's timsTOF mass spectrometers, stored in analysis.tdf and analysis.tdf_bin files inside .d folders. Frames can be selected by number or by retention time, and each peak is reported with its scan, time-of-flight index, intensity, mass-to-charge ratio and inverse ion mobility. Mass-to-charge ratios and ion mobilities are computed with built-in open-source calibration, or optionally with Bruker's proprietary library. See Łącki et al. (2021) .
- 22 October 2026
- Published by The R Foundation
opentimsr: An Open-Source Loader for Bruker's timsTOF Data Files
Reads raw data from Bruker's timsTOF mass spectrometers, stored in analysis.tdf and analysis.tdf_bin files inside .d folders. Frames can be selected by number or by retention time, and each peak is reported with its scan, time-of-flight index, intensity, mass-to-charge ratio and inverse ion mobility. Mass-to-charge ratios and ion mobilities are computed with built-in open-source calibration, or optionally with Bruker's proprietary library. See Łącki et al. (2021) .
- Open Access
Discriminative learning of substitution matrices and gap penalties for pairwise alignment of biological sequences
Pairwise alignment scores are used to classify pairs of sequences in many areas of bioinformatics, including homology search, predicting interactions, or read mapping. The relative scores of different pairs strongly depend on the choice of a substitution matrix and gap penalties, but the existing approaches for the estimation of these parameters do not directly optimize them for the task of classification. In this work, we present DiscrimAlign, a statistical model for discriminative learning of substitution matrices and gap penalties from a dataset of positive and negative pairs of unaligned biological sequences. The model links the alignment score of a sequence pair with the associated binary label through a logistic function and learns the parameters by likelihood maximization. We analyze theoretical properties of the model, derive and implement a learning procedure, study its performance in simulated experiments, and apply it to predict microRNA-target interactions. We show that sequence alignment with discriminative substitution matrices and gap penalties predicts the interactions comparably to state-of-the-art neural network classifiers while being more interpretable. An implementation of the model and reproducibility workflows are available at https://github.com/BioGeMT/DiscrimAlign.
Open AccessDiscriminative learning of substitution matrices and gap penalties for pairwise alignment of biological sequences
Pairwise alignment scores are used to classify pairs of sequences in many areas of bioinformatics, including homology search, predicting interactions, or read mapping. The relative scores of different pairs strongly depend on the choice of a substitution matrix and gap penalties, but the existing approaches for the estimation of these parameters do not directly optimize them for the task of classification. In this work, we present DiscrimAlign, a statistical model for discriminative learning of substitution matrices and gap penalties from a dataset of positive and negative pairs of unaligned biological sequences. The model links the alignment score of a sequence pair with the associated binary label through a logistic function and learns the parameters by likelihood maximization. We analyze theoretical properties of the model, derive and implement a learning procedure, study its performance in simulated experiments, and apply it to predict microRNA-target interactions. We show that sequence alignment with discriminative substitution matrices and gap penalties predicts the interactions comparably to state-of-the-art neural network classifiers while being more interpretable. An implementation of the model and reproducibility workflows are available at https://github.com/BioGeMT/DiscrimAlign.
- Open Access
- 4 May 2026
- Published by Oxford University Press (OUP) in Briefings in Bioinformatics
- Vol. 27 (3)
WNetAlign: fast and accurate spectra alignment using truncated Wasserstein distance and network simplex
Liquid chromatography–mass spectrometry (LC–MS) and nuclear magnetic resonance (NMR) spectroscopy are complementary analytical techniques widely used in proteomics, metabolomics, and structural biology. Both generate high-dimensional, noisy spectra where overlapping peaks complicate interpretation. LC–MS relies on retention time (RT) separation before mass analysis, while multidimensional NMR spreads information across chemical-shift axes to reduce congestion. However, comparative or replicate experiments often introduce RT shifts in LC–MS or frequency shifts in NMR, hindering accurate matching of corresponding features. In some experiments, such as variable-temperature NMR, the shifts are intentionally triggered, and frequency tracking provides important information. In any case, a robust, scalable alignment across runs is critical for reliable compound identification, quantification, and structural analysis. We propose a truncated Wasserstein distance-based algorithm for aligning LC–MS and NMR spectra. By constraining maximum transport distance and formulating alignment as a minimum-cost flow problem solved via the Network Simplex algorithm, our method accelerates computation, suppresses spurious matches, and improves robustness to noise. On benchmark LC–MS datasets, it achieved 0.97 precision, 0.96 recall, and a 0.6-s runtime, outperforming OpenMS and DeepRTAlign tools. For NMR data, the algorithm proved effective in 2D, 4D, and even 7D analyses. The algorithm is implemented in wnetalign with supporting modules wnet and pylmcf, available on PyPI and GitHub under permissive licenses: https://github.com/michalsta/pylmcf, https://github.com/michalsta/wnet, https://github.com/michalsta/wnetalign.
- 4 May 2026
- Published by Oxford University Press (OUP) in Briefings in Bioinformatics
- Vol. 27 (3)
Open AccessWNetAlign: fast and accurate spectra alignment using truncated Wasserstein distance and network simplex
Liquid chromatography–mass spectrometry (LC–MS) and nuclear magnetic resonance (NMR) spectroscopy are complementary analytical techniques widely used in proteomics, metabolomics, and structural biology. Both generate high-dimensional, noisy spectra where overlapping peaks complicate interpretation. LC–MS relies on retention time (RT) separation before mass analysis, while multidimensional NMR spreads information across chemical-shift axes to reduce congestion. However, comparative or replicate experiments often introduce RT shifts in LC–MS or frequency shifts in NMR, hindering accurate matching of corresponding features. In some experiments, such as variable-temperature NMR, the shifts are intentionally triggered, and frequency tracking provides important information. In any case, a robust, scalable alignment across runs is critical for reliable compound identification, quantification, and structural analysis. We propose a truncated Wasserstein distance-based algorithm for aligning LC–MS and NMR spectra. By constraining maximum transport distance and formulating alignment as a minimum-cost flow problem solved via the Network Simplex algorithm, our method accelerates computation, suppresses spurious matches, and improves robustness to noise. On benchmark LC–MS datasets, it achieved 0.97 precision, 0.96 recall, and a 0.6-s runtime, outperforming OpenMS and DeepRTAlign tools. For NMR data, the algorithm proved effective in 2D, 4D, and even 7D analyses. The algorithm is implemented in wnetalign with supporting modules wnet and pylmcf, available on PyPI and GitHub under permissive licenses: https://github.com/michalsta/pylmcf, https://github.com/michalsta/wnet, https://github.com/michalsta/wnetalign.
- Open Access
- 28 May 2024
- Published by American Chemical Society (ACS) in Analytical Chemistry
- Vol. 96 (23) , 9343-9352
MIND4OLIGOS: Determining the Monoisotopic Mass of Oligonucleotides Observed in High-Resolution Mass Spectrometry
Oligonucleotide therapeutics have emerged as an important class of drugs offering targeted therapeutic strategies that complement traditional modalities, such as monoclonal antibodies and small molecules. Their unique ability to precisely modulate gene expression makes them vital for addressing previously undruggable targets. A critical aspect of developing these therapies is characterizing their molecular composition accurately. This includes determining the monoisotopic mass of oligonucleotides, which is essential for identifying impurities, degradants, and modifications that can affect the drug efficacy and safety. Mass spectrometry (MS) plays a pivotal role in this process, yet the accurate interpretation of complex mass spectra remains challenging, especially for large molecules, where the monoisotopic peak is often undetectable. To address this issue, we have adapted the MIND algorithm, originally developed for top-down proteomics, for use with oligonucleotide data. This adaptation allows for the prediction of monoisotopic mass from the more readily detectable, most-abundant peak mass, enhancing the ability to annotate complex spectra of oligonucleotides. Our comprehensive validation of this modified algorithm on both in silico and real-world oligonucleotide data sets has demonstrated its effectiveness and reliability. To facilitate wider adoption of this advanced analytical technique, we have encapsulated the enhanced MIND algorithm in a user-friendly Shiny application. This online platform simplifies the process of annotating complex oligonucleotide spectra, making advanced mass spectrometry analysis accessible to researchers and drug developers. The application is available at https://valkenborg-lab.shinyapps.io/mind4oligos/.
- 28 May 2024
- Published by American Chemical Society (ACS) in Analytical Chemistry
- Vol. 96 (23) , 9343-9352
Open AccessMIND4OLIGOS: Determining the Monoisotopic Mass of Oligonucleotides Observed in High-Resolution Mass Spectrometry
Oligonucleotide therapeutics have emerged as an important class of drugs offering targeted therapeutic strategies that complement traditional modalities, such as monoclonal antibodies and small molecules. Their unique ability to precisely modulate gene expression makes them vital for addressing previously undruggable targets. A critical aspect of developing these therapies is characterizing their molecular composition accurately. This includes determining the monoisotopic mass of oligonucleotides, which is essential for identifying impurities, degradants, and modifications that can affect the drug efficacy and safety. Mass spectrometry (MS) plays a pivotal role in this process, yet the accurate interpretation of complex mass spectra remains challenging, especially for large molecules, where the monoisotopic peak is often undetectable. To address this issue, we have adapted the MIND algorithm, originally developed for top-down proteomics, for use with oligonucleotide data. This adaptation allows for the prediction of monoisotopic mass from the more readily detectable, most-abundant peak mass, enhancing the ability to annotate complex spectra of oligonucleotides. Our comprehensive validation of this modified algorithm on both in silico and real-world oligonucleotide data sets has demonstrated its effectiveness and reliability. To facilitate wider adoption of this advanced analytical technique, we have encapsulated the enhanced MIND algorithm in a user-friendly Shiny application. This online platform simplifies the process of annotating complex oligonucleotide spectra, making advanced mass spectrometry analysis accessible to researchers and drug developers. The application is available at https://valkenborg-lab.shinyapps.io/mind4oligos/.
- Top Cited Papers
- 16 February 2024
- Published by Springer Nature in Nature Methods
- Vol. 21 (3) , 365-367
OpenMS 3 enables reproducible analysis of large-scale mass spectrometry data
- 16 February 2024
- Published by Springer Nature in Nature Methods
- Vol. 21 (3) , 365-367
Top Cited PapersOpenMS 3 enables reproducible analysis of large-scale mass spectrometry data
- Open Access
midiaPASEF maximizes information content in data-independent acquisition proteomics
Data-independent acquisition (DIA) approaches provide comprehensive records of all detectable pre-cursor and fragment ions. Here we introduce midiaPASEF, a novel DIA scan mode using mobility-specific micro-encoding of overlapping quadrupole windows to optimally cover the ion population in the ion mobility-mass to charge plane. Using overlapping ion mobility-encoded quadrupole windows, midiaPASEF maximizes information content in DIA acquisitions which enables the determination of the precursor m/z of each fragment ion with a precision of less than 2 Th. The Snakemake-based MIDIAID pipeline integrates algorithms for multidimensional peak detection and for machine-learning-based classification of precursor-fragment relationships. The MIDIAID pipeline enables fully automated processing and multidimensional deconvolution of midia-PASEF files and exports highly specific DDA-like MSMS spectra which are suitable for de novo sequencing and can be searched directly with established tools including PEAKS, FragPipe and Mascot. midiaPASEF acquisition identifies over 40 unique peptides per second and provides powerful library-free DIA analyses including phosphopeptidome and immunopeptidome samples.
Open AccessmidiaPASEF maximizes information content in data-independent acquisition proteomics
Data-independent acquisition (DIA) approaches provide comprehensive records of all detectable pre-cursor and fragment ions. Here we introduce midiaPASEF, a novel DIA scan mode using mobility-specific micro-encoding of overlapping quadrupole windows to optimally cover the ion population in the ion mobility-mass to charge plane. Using overlapping ion mobility-encoded quadrupole windows, midiaPASEF maximizes information content in DIA acquisitions which enables the determination of the precursor m/z of each fragment ion with a precision of less than 2 Th. The Snakemake-based MIDIAID pipeline integrates algorithms for multidimensional peak detection and for machine-learning-based classification of precursor-fragment relationships. The MIDIAID pipeline enables fully automated processing and multidimensional deconvolution of midia-PASEF files and exports highly specific DDA-like MSMS spectra which are suitable for de novo sequencing and can be searched directly with established tools including PEAKS, FragPipe and Mascot. midiaPASEF acquisition identifies over 40 unique peptides per second and provides powerful library-free DIA analyses including phosphopeptidome and immunopeptidome samples.
- Open Access
- 12 October 2022
- Published by American Chemical Society (ACS) in Journal of the American Society for Mass Spectrometry
- Vol. 33 (11) , 2063-2069
Envemind: Accurate Monoisotopic Mass Determination Based On Isotopic Envelope
Nowadays, monoisotopic mass is used as an important feature in top-down proteomics. Knowing the exact monoisotopic mass is helpfu l for precise and quick protein identification in large protein databases. Howe v e r , only in spectra of small molecules the monoisotopic peak is visible. For bigger molecules like proteins, it is hidden in noise or undetected at all, and therefore its position has to be predicted. By improving the prediction of the peak, we contribute to a more accurate identification of molecules, which is crucial in fields such as chemistry and medicine. In this work, we present the envemind algorit h m , which is a two-step procedure to predict monoisotopic masses of proteins. The prediction is based on an isotopic envelope. Therefore, envemind is dedicated to spectra where we are able to resolve the one dalton separated isotopic variants. Furthermore, only single-molecule spectra are allowed, that is, spectra that do not req u i r e prior deconvolution. The algorit h m deals with the problem of off-by-one dalton errors, which are common in monoisotopic mass prediction. A novel aspect of this work is a mathematical exploration of the space of molecules, where we equate chemical formulas and their theoretical spectrum. Since the space of molecules consists of all possible chemical formulas, this approach is not limited to known substances only. This makes optimization processes faster and enables to approximate theoretical spectrum for a given experimental one. The algorithm is available as a Python package envemind on our GitHub page https://github.com/PiotrRadzinski/envemind .
- 12 October 2022
- Published by American Chemical Society (ACS) in Journal of the American Society for Mass Spectrometry
- Vol. 33 (11) , 2063-2069
Open AccessEnvemind: Accurate Monoisotopic Mass Determination Based On Isotopic Envelope
Nowadays, monoisotopic mass is used as an important feature in top-down proteomics. Knowing the exact monoisotopic mass is helpfu l for precise and quick protein identification in large protein databases. Howe v e r , only in spectra of small molecules the monoisotopic peak is visible. For bigger molecules like proteins, it is hidden in noise or undetected at all, and therefore its position has to be predicted. By improving the prediction of the peak, we contribute to a more accurate identification of molecules, which is crucial in fields such as chemistry and medicine. In this work, we present the envemind algorit h m , which is a two-step procedure to predict monoisotopic masses of proteins. The prediction is based on an isotopic envelope. Therefore, envemind is dedicated to spectra where we are able to resolve the one dalton separated isotopic variants. Furthermore, only single-molecule spectra are allowed, that is, spectra that do not req u i r e prior deconvolution. The algorit h m deals with the problem of off-by-one dalton errors, which are common in monoisotopic mass prediction. A novel aspect of this work is a mathematical exploration of the space of molecules, where we equate chemical formulas and their theoretical spectrum. Since the space of molecules consists of all possible chemical formulas, this approach is not limited to known substances only. This makes optimization processes faster and enables to approximate theoretical spectrum for a given experimental one. The algorithm is available as a Python package envemind on our GitHub page https://github.com/PiotrRadzinski/envemind .
- 30 September 2020
- Published by Wiley in Rapid Communications in Mass Spectrometry
- Vol. 39 (S1)
Masserstein: Linear regression of mass spectra by optimal transport
Rationale The linear regression of mass spectra is a computational problem defined as fitting a linear combination of reference spectra to an experimental one. It is typically used to estimate the relative quantities of selected ions. In this work, we study this problem in an abstract setting, in order to develop new approaches applicable to a diverse range of experiments. Methods In order to overcome the sensitivity of the ordinary least squares regression to measurement inaccuracies, we base our methods on a non‐conventional spectral dissimilarity measure, known as the Wasserstein or the Earth Mover's distance. This distance is based on the notion of the cost of transporting signal between mass spectra, which renders it naturally robust to measurement inaccuracies in the mass domain. Results Using a data set of 200 mass spectra, we demonstrate that our approach is capable of accurate estimation of ion proportions without extensive pre‐processing of spectra required by other methods. The conclusions are further substantiated using data sets simulated in a way that mimics most of the measurement inaccuracies occurring in real experiments. Conclusions We have developed a linear regression algorithm based on the notion of the cost of transporting signal between spectra. Our implementation is available in a Python 3 package called masserstein, freely available at https://github.com/mciach/masserstein.
- 30 September 2020
- Published by Wiley in Rapid Communications in Mass Spectrometry
- Vol. 39 (S1)
Masserstein: Linear regression of mass spectra by optimal transport
Rationale The linear regression of mass spectra is a computational problem defined as fitting a linear combination of reference spectra to an experimental one. It is typically used to estimate the relative quantities of selected ions. In this work, we study this problem in an abstract setting, in order to develop new approaches applicable to a diverse range of experiments. Methods In order to overcome the sensitivity of the ordinary least squares regression to measurement inaccuracies, we base our methods on a non‐conventional spectral dissimilarity measure, known as the Wasserstein or the Earth Mover's distance. This distance is based on the notion of the cost of transporting signal between mass spectra, which renders it naturally robust to measurement inaccuracies in the mass domain. Results Using a data set of 200 mass spectra, we demonstrate that our approach is capable of accurate estimation of ion proportions without extensive pre‐processing of spectra required by other methods. The conclusions are further substantiated using data sets simulated in a way that mimics most of the measurement inaccuracies occurring in real experiments. Conclusions We have developed a linear regression algorithm based on the notion of the cost of transporting signal between spectra. Our implementation is available in a Python 3 package called masserstein, freely available at https://github.com/mciach/masserstein.
- Open Access
Masserstein: robust linear deconvolution by optimal transport
A common theme in many applications of computational mass spectrometry is fitting a linear combination of reference spectra to an experimental one in order to estimate the quantities of different ions, potentially with overlapping isotopic envelopes. In this work, we study this procedure in an abstract setting, in order to develop new approaches applicable to a diverse range of experiments. We introduce an application of a new spectral dissimilarity measure, known in other fields as the Wasserstein or the Earth Mover’s distance, in order to overcome the sensitivity of ordinary linear regression to measurement inaccuracies. Usinga a data set of 200 mass spectra, we demonstrate that our approach is capable of accurate estimation of ion proportions without extensive pre-processing required for state-of-the-art methods. The conclusions are further substantiated using data sets simulated in a way that mimics most of the measurement inaccuracies occurring in real experiments. We have implemented our methods in a Python 3 package, freely available at https://github.com/mciach/masserstein .
Open AccessMasserstein: robust linear deconvolution by optimal transport
A common theme in many applications of computational mass spectrometry is fitting a linear combination of reference spectra to an experimental one in order to estimate the quantities of different ions, potentially with overlapping isotopic envelopes. In this work, we study this procedure in an abstract setting, in order to develop new approaches applicable to a diverse range of experiments. We introduce an application of a new spectral dissimilarity measure, known in other fields as the Wasserstein or the Earth Mover’s distance, in order to overcome the sensitivity of ordinary linear regression to measurement inaccuracies. Usinga a data set of 200 mass spectra, we demonstrate that our approach is capable of accurate estimation of ion proportions without extensive pre-processing required for state-of-the-art methods. The conclusions are further substantiated using data sets simulated in a way that mimics most of the measurement inaccuracies occurring in real experiments. We have implemented our methods in a Python 3 package, freely available at https://github.com/mciach/masserstein .