Time: Fridays 10.00AM. Location: IPI PAN building, 3rd floor, room 334.
Seminars 2022
May 6th 2022
A talk given by Michal J. Dabrowski entitled: "Unveiling transcriptomic effects of tuberculosis vaccines". Summary provided by the author: In the lab experiment two separate groups of Rhesus Macaques have been tested, using two different tuberculosis vaccines and to compare their
response the third control group was used as well. Each group consists of 12
animals and their gene expression data was obtained in 10 different time
points. Although the analysis is extremely difficult because of observed high
mortality among all these groups and the number of animals in the decision
classes is highly limited. To deal with the problems two independent
approaches were used: the MCFS feature selection compared to the
construction of linear models provided by DESeq2.
April 29th 2022
A talk given by Mateusz Jakiel entitled: "Markers of disease and cell type in Alzheimer's disease from a scRNA-seq perspective". Summary provided by the author: scRNA-seq is a method that allows researchers to examine the changes in gene expression from a resolution not possible before. That is why, by comparing the analysis results of multiple datasets, we can try to find potentially novel markers of diseases such as Alzheimer's disease with more precision than with the analysis of bulk RNA-seq. We can also extrapolate interesting biomarkers which are key to correctly annotating the cell types in the dataset. This seminar will focus on both of these types of biomarkers and the methods used to achieve them.
Seminars 2021
November 26th 2021
A talk given by Michał Dramiński entitled: "Variable Selection with Knockoffs". Summary provided by the author: Feature selection is an important task in modern analysis of molecular data. In many problems we deal with thousands of features/variables and the problem is to select informative set excluding false positives. The talk is focused on presentation of a new method of evaluation of FDR based on knockoff features as well as comparison of the method and the results to MCFS-ID algorithm.
November 5th 2021
A talk given by Agata Dziedzic entitled: "Epigenetic patterns of enhancers specific for glioma samples of different grades". Summary provided by the author: Gliomas are the most common primary tumors of the central nervous system, and are also among the deadliest types of cancer. There are evidences that epigenetic changes contribute to modulating gene expression in glioma. The goal of this study was to compared gene regulation mechanisms via enhancers in pilocytic astrocytoma (WHO grade I, PA) to higher grade glioma samples (WHO grade II-IV, HGG). Locations of active enhancers were previously determined by peaks of acetylated H3K27 in non-promoter regions. We analysed DNA methylation levels in and performed transcription factors motif search in 239 grade-specific enhancers as well as in 275 additional, non grade-specific enhancers. Global DNA methylation level was higher in HGG samples comparing to PA samples. Methylation pattern of 14 motifs was confirmed to be consequently hypermethylated in HGG compared to PA samples (chi-square test). We also report set of single CpG sites that are differentially methylated between HGG and PA samples together with their locations in transcription factor motifs. These results indicate specific transcription factor motifs whose methylation may have an influence on regulation of gene expression and therefore contribute to gliomagenesis.
July 9th 2021
A talk given by Damian Loska entitled: "SNPs in Symfonia Project". Summary provided by the author: The samples from Symfonia Project were sequenced using distinct technologies: DNA-seq, Methyl-seq. Each panel covered distinct areas of the genome, but there was a small overlap. Calling SNPs from DNA-seq nowadays is a relatively easy task, however obtaining variants from Methyl-seq is more challenging. We tried to utilize the SNPs data from both experiments in order to find correlation with gene expressions. Firstly we assessed how many SNPs we obtain from each experiment in the common genomic regions (overlapped panels). Further we performed statistical tests to find the association between 0/0, 0/1, 1/1 genotypes in the promoter area and gene expression levels. The results are visualised using rShiny.
May 14th 2021
A talk given by Mateusz Jakiel & Michał J. Dąbrowski entitled: "Modulation of astrocyte reactivity improves functional deficits in mouse models of Alzheimer's disease - analysis of data and summary". Summary provided by the authors: The goal of this seminar was to elaborate on certain biological aspects of the paper, mainly the types of model organisms used for testing and what characterized them, so as to make it clear for everybody who wishes to look at the data and analyse it for themselves. Each transgenic change in every mouse model used in this study was described and explained in detail. The second aim of this seminar was to present the results of an analysis of the data from the publication: "Mapping chromatin accessibility and active regulatory elements reveals new pathological mechanisms in human gliomas" [Paper #NCOMMS-20-13682B].
April 2nd 2021
A talk given by Michał J. Dąbrowski entitled: "DNA methylation analysis of regulatory elements of TGF- β pathway related genes with differential expression: in terms of whole promoters as well as selected short regions which are binding sites for transcription factors". Summary provided by the author: Transforming growth factor-beta (TGF-β) is a multifunctional cytokine that regulates cell proliferation, cell differentiation, apoptosis, extracellular matrix development, and has immunomodulatory properties. Many of these mechanisms are closely related, and any disturbances in them can lead to diseases, including connective tissue diseases (fibrosis, cardiovascular disease) and cancer. In cells susceptible to TGF-β stimulation, its action causes cell cycle arrest in the late G1 phase, induction of apoptosis, and the expression of cell adhesion molecules including fibronectin, laminin, collagen and integrins. However, after cell tumor transformation, autocrine or paracrine of TGF-β enhances tumorigenesis progression, stimulating angiogenesis, invasiveness and proliferation. In this project, we focused on verifying whether different DNA methylation levels within regulatory regions of almost one hundred genes related to the TGF-β signalling pathway, are modulating those gene expression. As a model we used glioma tumor samples from The Cancer Genome Atlas as well as Symfonia III project. The later consists data from high-throughput experiments: ATAC-seq, H3K4me3, H3K27ac ChIP-seqs, DNAse-seq, DNA bisulfite sequencing and RNA-seq acquired from glioma bulk tumour samples of various WHO grades.
March 26th 2021
A talk given by Damian Loska entitled: "Microbiome 16S rRNA: data processing workflow and statistical analysis". Summary provided by the author: Microorganisms play an important role in every aspect of the world's life: shaping global and local natural environment, as well as influencing the well-being of all higher organisms including humans. Metagenomics since years has been a robust method for microorganisms detection and identification, however still it remains too expensive for multi sample studies. 16S rRNA sequencing is a fast and cheap alternative to metagenomics. Even though the resolution of 16S rRNA allows to discriminate microorganisms up to genus level, still the method is good enough to identify differences in groups of patients. Nonetheless, the statistical analysis of such full-of-noise data remains challenging. Here we present a common approach to obtain reliable information from 16S rRNA sequencing data.
March 12th 2021
A talk given by Bartosz Wojtaś entitled: "REST transcription factor holds the balance between the invasion and cell differentiation in IDH-mutant and IDH-wild type gliomas". Summary provided by the author: REST transcription factor is a known repressor of neuronal genes in non-neuronal cells. Moreover in the oncology context it has been recognized as gene expression activator of many genes, recognized as oncogenes. REST is known to be affected and to shape local epigenetic environment. Therefore we have studied REST binding in the context of IDH-phenotype. Mutation in IDH1 or IDH2 genes, causing so called IDH-phenotype produce a massive change in the epigenetic landscape. We have used siRNA technology to silence REST gene and we have used global chromatin immunoprecipitation profiles of REST to define differences between IDH mutants and IDH wild-type cell lines. Lastly we have verified our findings in the tumor samples from The Cancer Genome Atlas repository and on own cohort of tumor samples. We have found that REST controls the shift between migration and cell differentiation and its effect is different, depending on IDH mutation status.
March 5th 2021
A talk given by Agata Dziedzic entitled: "How IDH1 mutation affects glioma - the most common type of primary brain tumor". Summary provided by the author: Glioma is the most common type of primary malignant brain tumor. As with many cancers, the causes of glioma are complex and, although extensively studied by research teams around the world, are still unclear. One of the frequently mentioned genetic aberrations underlying the disease is a mutation in the IDH1 gene. This mutation is believed to influence the dynamics of disease development and more and more research is being done on it, but the overall impact of IDH1 gene mutation on the oncogenesis process is not yet fully understood. Our research addresses this issue. The primary research material consisted of glioma samples taken from patients during surgery. These samples were subjected to a series of laboratory experiments which provided data on several different layers of genetic information. We used the obtained data to perform a comparative analysis between samples with different IDH1 gene status: IDH1-mutants were compared to so-called IDH1-wild type. We obtained results that indicate specific consequences of the IDH1 mutation and suggest how the cascade of successive events triggered by this mutation may influence the development of the disease. The IDH1 mutation causes a domino effect, disrupting the regulation of subsequent genes - we show how it can proceed and which genes are ultimately disturbed. We hope that by reporting the genetic differences between IDH1-mutants and wild-type IDH1, we will contribute to our understanding of the complex implications of the IDH1 mutation.
January 22nd 2021
A talk given by Michał J. Dąbrowski entitled: "The role of differential DNA methylation levels in expression regulation of genes related to TGF- β pathway"
Seminars 2020
December 18th 2020
A talk given by Agata Dziedzic entitled: "Methylation pattern in acitve enhancers in pilocytic astrocytoma and Higher Grade Glioma samples". Summary provided by the author: A technical update on progress of work on enhancer regions in glioma samples in Symfonia 3 project.
December 11th 2020
A talk given by Marta Jardanowska entitled: "Unveiling significant molecular markers in breast cancer prediction". Summary provided by the author: Our study aimed to uncover significant molecular features and interdependencies that are significant for breast cancer prediction. The Monte Carlo Feature Selection and Interdependence Discovery algorithm was applied to select the most significant out of high dimensional molecular data from The Cancer Genome Atlas (TCGA). We identified 2006 DNA methylation sites relevant in cancer prediction, which have been divided depending on the methylation level. For important DNA methylations and their surroundings +/-20bp, the search of transcription factors (TFs) was performed. Using this data, we have identified interdependencies between TF target genes, genes encoding TFs and DNA methylations sites. Selected DNA methylation sites may affect TF binding affinity resulting in target gene expression changes.
December 4th 2020
A talk given by Mateusz Jakiel entitled: "Bacterial speech and how AI could analyze it". Summary provided by the author: Quorum sensing is a complicated mechanism by which bacteria communicate with each other using different types of signaling molecules. Because of the large variety of these molecules it is extremely difficult and time consuming to research how they might interact with cells other than bacteria. Integrating artificial intelligence in this process might be the next step in searching for theoretical uses of the molecules in creating new cures for some diseases like muscle atrophy.
November 20th 2020
A talk given by Paweł Teisseyre entitled: "Introduction to positive-unlabelled learning (PU learning)". Summary provided by the author: The aim of the presentaion is to describe a problem of positive-unlabelled classification, which nowadays is a hot topic in machine learning. In traditional, binary classification we try to build a classification model which distinguish between positive and negative classes. In PU learning, there are only positive examples and unlabelled examples in training data. Unlabelled examples can be either positive or negative. For example, in medical applications, the unlabelled class may correspond to undiagnosed diseases. The absence of a diagnosis does not mean that the patient does not have a disease in question. In the talk I will discuss basic methods used in PU learning. In particular, I will show how to adapt existing simple models (e.g. logistic regression) to the case of PU data.
October 30th 2020
A talk given by Małgorzata Perycz entitled: "Possible effectors of REST in glioma". Summary provided by the author: REST is a transcription factor and repressor that is implicated in both protective and cell death mechanisms. Clasically, it represses neural genes in non neuronal cells. In glioma, lower REST levels correlate with longer survival of the patient. In this talk I show the results obtained in glioma cell line models in which REST was knocked down. The data is presented in the context of IDH1 mutation, which is known to alter the epigenetic landscape of the cell and which is an early mutation in development of glioma.
October 8th 2020
A talk given by Michal Draminski entitled: "Good practices in R programing based on CytoMeth example". Summary provided by the author: During the talk a set of good practices in design and implementation of R tools will be provided. CytoMeth is a R tool that evolved from set of loose R scripts into GitHub project that can be used by any user on his own data. The speaker will also share some tips and tricks that he used during implementation to speed up the processing of huge data (e.g. data.frames) in CytoMeth.
September 25th 2020
A talk given by Agata Dziedzic entitled: "Regulation of gene expression via enhancers methylation in pilocytic astrocytoma and higher grade glioma samples". Summary provided by the author: Gliomas are most common tumors of the central nervous system, and are also among the most deadly types of cancer. There are evidences that epigenetic changes contribute to modulating gene expression in glioma. In this study we aimed to compare gene regulation mechanisms via enhancers in pilocytic astrocytoma (PA) samples to Higher Grade Glioma (HGG) samples (diffuse astrocytoma and glioblastoma histopatological grades). Results showed that HGG enhancers had lower frequency of guanine and cytosine nucleotides (GC content) then PA specific enhancers and higher global DNA methylation level. Genes targeted by identified enhancers were assigned based on enhancer location combined with Hi-C data – PA enhancers targeted from 1 up to 15 target genes (TG) and HGG from 1 to 6 TG. Chi-squared test revealed 14 motifs whose methylation pattern was consequently different between its occurrences in PA and HGG samples. This results indicate specific TF motifs whose methylation may have an influence on regulation of TG expression.
September 18th 2020
A talk given by Mateusz Rezler entitled: "Synthesis, structure analysis and assessment of biological activity of new aromatic diamidines". Summary provided by the author: An extract of my PhD thesis that covers project of new compounds, organic preparation, spatial structure analysis, evaluation of biological activity, and modelling of interactions with postulated molecular target. Research was focused on pentamidine analogs that target minor groove of DNA and inhibit the growth of Pneumocystis fungi. At the end of the speech, the results and future perspectives were discussed.
September 3rd 2020
A talk given by Michał J. Dąbrowski entitled: "Detection of differentially methylated short DNA sequences within MGMT promoter". Summary provided by the author: DNA methylation level of the promoter of MGMT gene is a well-known molecular marker associated with glioma patient’s prognosis. Here it will be shown how the use of “short sequence search” function, which will be implemented in the CytoMeth tool, may help in unveiling the association of DNA methylation and MGMTgene expression.
July 31st 2020
A talk given by Marcin Łapiński entitled: "Potential TF – TFBS methylation control networks of genes important in breast cancer prediction". Summary provided by the author: Presentation of search of potental control mechanisms of gene expressions which are important in cancer development or cancer prediction. Hypotized mechanism employ transcription factors and methylation of cytosines within transcription factor binding sites. Analysis utilized MCFS-ID, bedtools and PWMEnrich tools.
June 4th 2020
A talk given by Mateusz Rezler entitled: "High‑throughput prediction of DNA shape features and its application to the analysis of differentially methylated CpG sites". Summary provided by the author: Information about the DNA sequence and shape is complementary in determining protein-DNA binding specificity. This seminar focuses on the shape and its efficient prediction by presenting examples of the state of the art solution of this problem (including methylated DNA).
May 21th 2020
A talk given by Marta Jardanowska entitled: "Identification and analysis of transcription factor motifs in REST ChIP-seq peaks for activated and repressed genes". Summary provided by the author: In collaboration with Nencki Institute, we started to analyse REST ChIP-seq peaks for activated and repressed genes. In my talk, I will present the overall results of TF motifs analysis such as TF motif search, functional analysis, clustering and representation of TF families. Additionally, I will show you my interesting discovery of co-occurrence of other TF motif with REST.
May 4th 2020
A talk given by Marcin Łapiński entitled: "Generative architectures of neural networks: deep dream, GAN and autoencoders". Summary provided by the author: An introduction into generative neural models. Presentation of mathematical basis, inner architecture, learning algoritms, properties and example applications of convolutional neural networks architectures listed in the title of this talk.
April 2nd 2020
A talk given by Agata Dziedzic entitled: "Enhancers common for three grades of glioma samples - Comparison of DNA methylations patterns and study of its influence on gene expression". Summary provided by the author: Our aim was to study molecular differences between pilocytic astrocytoma (grade I; n = 7) and higher grade glioma (glioblastoma (grade IV; n = 6) and diffuse astrocytoma (grade II/III; n = 4) samples. We used H3K27ac Chip-seq data to define enhancer regions active both in PA in higher grade glioma and then RNA-seq and Bisulphite-seq data to select specific CpG sites in transcription factor motifs responsible for changes of its transcription factors binding affinity. Here we report 7 deferentially methylated CpG sites that possibly influence expression level of target genes.
March 12th 2020
A talk given by Michał Dąbrowski entitled: "Glioma samples methylation levels within TF binding sites predicted using glioblastoma cell line ATAC-seq data". Summary provided by the author: BMO is a tool that performs prediction of transcription factor binding sites (TFBS) based on ATAC-seq data. Here, we predicted TFBS within glioblastoma cell line LN18. Using known motifs from HOCOMOCO database and LN18 ATAC-seq data the putative binding sites were assigned. With use of glioma tumour methylomes it was verified whether any motif has changed methylation pattern globally - within all its binding sites when compared between different glioma WHO grades.
February 27th 2020
A talk given by Ilona Grabowicz entitled: "Application of neural networks to solve some biological problems". Summary provided by the author: In my talk, I will present methodology and preliminary results of my work done in collaboration with Marcin Lapinski. We apply neural networks models to find characteristic motifs in the sequences of promoters and enhancers de-/activating genes involved in glioma initiation and/or progression. In the second project, we train the model to enable automated recognition of different plant species in the microscopic images.
January 30th 2020
A talk given by Agata Dziedzic entitled: "DNA methylation of transcription factor motifs within enhancers - glioma grades comparison". Summary provided by the author: DNA methylation within enhancers is known to affect gene expression. We studied enhancer regions to find molecular differences between high and low grade glioma samples. Data we used included 1) single methylation levels (Bisulphite conversion followed by sequencing), 2) gene expression levels (RNA-seq) and 3) transcription factor motifs found on enhancer sequences (ChIP-seq experiment for H3K27ac). As a result, we report a set of differentially methylated transcription factor motifs together with specific cytosines in those motifs whose methylation status affects the level of expression of the target gene. These specific changes in DNA could play a role in glioma development and are potential candidates for experimental validation.
Seminars 2019
December 12th 2019
A talk given by Marcin Łapiński entitled: "Relevant feature selection for Breast Cancer prediction using MCFS algorithm. Comparison to previous results. Bonus: Monte-Carlo error propagation model for MCFS". Summary provided by the author: In my talk, I will present results of the MCFS algorithm - relevant features in breast cancer prediction. I will discuss difference with results previously. Additionally, I will present method of Monte Carlo error propagation simulation for MCFS with example of use.
November 21st 2019
A talk given by Marta Jardanowska entitled: "Identification of transcription factor motifs influenced by methylation level in breast cancer." Summary provided by the author: During the presentation, the author will present results for 612 sequences containing hypomethylated or hypermethylated cytosines significant for breast cancer prediction. Sequences were divided into two groups - hypo or hypermethylated. While this talk, the result of the TF motifs search, functional analysis, expression changes and correlation between methylation and expression levels will be presented. For PLAL1 protein, whose gene was identified by MCFS as the only one statistically significantly predicted breast cancer, the author will also present preliminary studies for 3D structure.
November 7th 2019
A talk given by Ilona Grabowicz entitled: "Chromatin-related epigenetic mechanisms regulate expression of genes in developing gliomas". Summary provided by the author: In my talk, I will present results of the analysis where I was investigating different mechanisms of regulation of gene expression in glioma samples. The mechanisms on which I will focus include epigenetic regulation with H3K27ac and H3K4me3 marks as well as enhancers.
October 24th 2019
A talk given by Michał Dąbrowski entitled: "Structuring the Symfonia methylation data in the context of the panel specificity as well as future global and locus specific DNA methylation analysis". Summary provided by the author: During the seminar the summary of the preliminary results of the DNA methylation data pre-processing will be shown. Because of the CytoMeth tool preparation, all Symfonia methylation data were re-computed and in consequence more samples can be included in the downstream analysis. Because of the small sample size and high variability in sequencing coverage of specific cytosines across samples, we try to find out the best solution how to detect locus specific methylation differences.
June 13th 2019
A talk given by Agata Dziedzic entitled: "CytoMeth - additional functionalities". Summary provided by the author: CytoMeth is going to be a tool for DNA methylation data analysis. Its main purpose is to take raw bisulphite (fastq) data as input and return beta value as result. On this seminar the author will present suggestions and ideas for additional functionalities of Cythometh, including figures with genomic annotation for batch samples input and broad quality control statistics together with biological explanation of it.
May 30th 2019
A presentation and discussion leaded by Michal Dramiński entitled "CytoMeth - The current beta version". Summary provided by the author: CytoMeth is a new upcoming tool implemented by ZBO team. The author is going to present the current beta version of the tool that compiles a set of open source software named in the Roche pipeline guidelines to perform SeqCap Epi data analysis.
May 23rd 2019
A discussion leaded by Michal J. Dąbrowski entitled "Discussion of the conference in CSH". Summary provided by the author: Because of my participation in the Cold Spring Harbor Laboratory conference about “The Biology of Genomes”, I would like to share my observations regarding computational biology in the field of molecular biology, newly acquired knowledge and the possibilities of applying newly discovered tools in our research.
May 16th 2019
A talk given by Marta Jardanowska entitled "Identification of transcription factor motifs in three different regions influenced by significant DNA methylations in breast cancer". Summary provided by the author: During the previous presentation, the author presented results in 1454 sequences. This time the author decided to use sequences +/- 20bp around all 2023 significant methylations that predict breast cancer and have been identified in the previous analysis. These sequences were divided into three groups depending on the localisation - in promoter regions, gene body regions or intergenic regions. While this talk, the result of the TF motifs search, functional analysis and expression changes will be presented.
May 9th 2019
A talk given by Agata Dziedzic entitled "DNA methylation of transcription factor motifs within promoters - glioma grades comparison". Summary provided by author: DNA methylation within promoter is knows to affect gene expression. This study aimed to find differentially methylated DNA transcription factor motifs within promoters and to combine obtained results with gene expression changes between glioma grades. Two different approaches were used to find differentially methylated motifs between grades: 1) single methylation site comparison and 2) mean methylation value of whole motif. As a result we raport differentially expressed genes whose expression might be regulated by DNA methylation in their promoters as well as set of specific transcription factors that would be involved in this process.
April 18th 2019
A discussion leaded by prof. Jacek Koronacki entitled "British Biologist Denis Noble Debunks Neo Darwinism".
April 4th 2019
A talk given by Marta Jardanowska entitled "Identification of transcription factor motifs influenced by significant DNA methylations in breast cancer". Summary provided by the author: Significant methylations that predict breast cancer have been identified in the previous analysis. In order to check if these methylations could affect any transcription factor (TF), the reference genome was used to determine the sequence +/- 20bp around these methylations. During this talk, the result of the TF motif search will be presented.
March 28th 2019
A talk given by dr Adam Filip entitled "BRCA - past, presence and future of the project". Summary: During the talk, author will present recent updates to the project together with future plans. In particular new analysis of methylations involved in predicting cancer will be presented.
March 21st 2019
A talk given by dr Michał J. Dąbrowski entitled "How to select meaningful regulatory regions – the case study of promoter methylations in various glioma tumours". Summary: DNA methylation patterns of genomic regulatory regions affect gene expression levels. In consequence genes that in normal tissue should be silenced are enhanced, and those which should be enhanced are repressed. The change of genes expression results in tumour progression and it seems that epigenetic changes play an important role in this process. Here we confirmed the association of promoter methylation level and gene expression and with use of relevant biological filter we selected the true positive examples.
March 14th 2019
A talk given by Ilona Grabowicz entitled "Analysis of methylation of cytosines localized within the TAD boundaries". Summary: During this seminar, the results of my latest analyzes on cytosine methylation within the TAD boundaries will be presented. The investigated topics included comparing levels of cytosines methylations depending on the: IDH mutation status, glioma WHO grade and CTCF binding regions. Moreover, it will be shown how methylation levels differ between TAD border and non-border cytosines and whether differentially methylated cytosines influence levels of co-regulation of gene expression within and accross neighboring TADs.
March 7th 2019
A talk given by dr Michał Dramiński entitled "Introduction to object-oriented programming (OOP)". Summary: During this talk object-oriented programming concept will be presented. OOP model is organized around objects rather than "actions" and data rather than logic. Together with the introduction to the paradigm, a real life and some low level (source code) examples will be shown.
February 28th 2019
A talk given by Agata Dziedzic entitled "An update on methylation of TFs motifs and its influence on gene expression - mean methylation instead of single CpG analysis". Summary: I kindly invite everybody who is interested in methylation of transcription factor motifs for my seminar that I prepared together with Marta J and Michał JD. Let's see which genes turned out to be statistically significant when using a different statistical method for detection and discuss how detected genes might explain biological processes in glioma. This time results were obtained by Wilcoxon test based on comparing different glioma grades. I look forward to your scientific criticism!.
February 21st 2019
A talk given by prof Jacek Koronacki entitled "From Modeling to Model-free Data Mining and Back". Summary: For a long time statistics had been a hypothesis-driven endeavor, tightly connected with modeling paradigm. Somewhere around 1975 the "Age of Tukey" followed under the banner of "Let tha data speak for themselves". Indeed, with more and more data at hand, modeling proved infeasible unless all too oversimplifying assumptions were made. As was expected, with Data Flood coming in, model-free Data Mining has brought much success to research and development, this being due to breathtaking increase in processing power and memory capacity of our PC's, not to mention HPC systems and cloud computing. However, by the very same token, our modeling abilities have also increased "almost infinitely". In this talk, we'll summarize briefly this process and make some remarks on the power of forward simulation studies (by examples from oncology and portfolio selection) and data-driven causal reasoning.
February 14th 2019
A talk given by Marta Jardanowska entitled "Identification of DNA motifs regulated by DNA methylation in glioma samples - preliminary study". Summary: Transcription factors (TF) regulate gene expression. TF binding affinity to its DNA motif might be affected by the methylation level of cytosines within that motif. The aim of this presentation will be showing the results of examining the methylation patterns of various TF motifs.
February 7th 2019
A talk given by dr Paweł Teisseyre entitled "A deeper look at two concepts of measuring gene-gene interactions: logistic regression and interaction information revisited". Summary: Detection of gene–gene interactions is one of the most important challenges in genome-wide case–control studies. Besides traditional logistic regression analysis, recently the entropy-based methods attracted a significant attention. Among entropy- based methods, interaction information is one of the most promising measures having many desirable properties. Although both logistic regression and interaction information have been used in several genome-wide association studies, the relationship between them has not been thoroughly investigated theoretically. The present paper attempts to fill this gap. We show that although certain connections between the two methods exist, in general they refer two different concepts of dependence and looking for interactions in those two senses leads to different approaches to interaction detection. The numerical experiments illustrate the theoretical findings indicating that interaction information and its modified version are more universal tools for detecting various types of interaction than logistic regression and linkage disequilibrium measures.
January 17th 2019
A talk given by dr Adam Filip entitled "Genomic features predicting presence of breast invasive carcinoma (BRCA)". Summary: During the talk results of analysis of breast cancer data from TCGA (The Cancer Genome Atlas) will be presented. The author will also present a set of mRNAs, methylations, miRNAs and their interplay that determines presence of cancer.
January 10th 2019
A talk given by dr Michał J. Dąbrowski entitled "Specific DNA methylation level of promoter regions detected for specific group of genes for pilocytic astrocytoma and glioblastoma tumor samples". Summary: DNA methylation is a well-known epigenetic element that regulates genes expression. It was shown that DNA methylation significantly affects glioma tumour development and is associated with patient’s prognosis. Here it was studied whether gene promoter various levels of methylation were correlated with expression of those genes. It was confirmed that a number of 74 genes differed in promoter level methylation as well as in gene expression between pilocytic astrocytoma samples and glioblastoma.
Seminars 2018
December 10th 2018
A talk given by Ilona Grabowicz entitled "Methylation of CpGs present within borders of Topologically Associating Domains". Summary: Within TAD borders there were found CpGs that can be methylated. The observed levels of methylation differed between glioma patients classified with different WHO grades and between patients carrying mutations in IDH genes and non-carriers.
December 6th 2018
A talk given by dr Michał Dramiński entitled "Classification of Symfonia patients based on models built on Cecarelli RNASeq data". Summary: This talk is about selection of the most significant genes for several target attributes (e.g. IDH mutant status, grade, survival, GBM subtype etc.) from Cecarelli RNASeq data, later building classification models and application of them on Symfonia RNASeq data. Selection of significant features is performed using MCFS-ID algorithm and classification method is based on application of committee of three different classifiers. Finally, a new set of descriptive features is obtained for Symfonia patients.
November 29th 2018
A talk given by Agata Dziedzic entitled "Differentially methylated sites and their influence on gene expression". Summary: We performed differentional methylation calling and found 209,260 differentially methylated sites for PA vs. GB and 174,383 differentially methylated sites for IDH wild-type vs. IDH mutant. Then we annotated those sites in respect to the genome parts (promoter, gene body, CpG islands). Additionally, to verify the differentially methylated sites functionality, we investigated expression (RNA-seq) of genes associated with those sites.
November 22nd 2018
A talk given by prof Jacek Koronacki entitled "AI: What's that, and how it relates to human enhancement and cyber immortality". Summary: The talk will consist of two parts. In the first, a study in progress which offers a novel approach to describe what AI is will be presented. By far, the main author of this novel approach is Andrzej Rabczenko, with the undersigned playing the role of a critical commentator of Andrzej’s work. In the second part, a hype on human enhancement, cyber immortality and the like endeavors will be critically (and with due hostility) assessed.
November 8th 2018
A talk given by Marta Jardanowska entitled "Motif analysis: detection of transcription factor motifs on promoter and enhancer sequences". Summary: The aim of this presentation will be introduction to motif searching in sequences in promoter and enhancer regions from patients diagnosed with glioma. The author will show common and differential motifs from PA and DA/GB samples using two different approaches of searching - de novo analysis and known motif search. Then the author will be showing some biological relations, especially protein families correlations.
October 25th 2018
A talk given by Adria Roura entitled "Somatic mutation history of glioblastoma patients with recurrent tumors". Summary: Glioblastoma (GBM, WHO grade IV) is a common and most lethal primary brain tumor which remains largely resistant to current therapies. Despite tumor resection and the following treatment of GBM patients with radiotherapy and chemotherapy, GBM frequently recurs. Dissection of the GBM genetics and understanding intrinsic mechanisms of tumor recurrence may lead to more targeted and effective treatments. Here we report the results of targeted next-generation sequencing of cancer- and epigenetics-related genes in 16 fresh frozen glioma samples of WHO grade IV, collected from Polish population. We employed a second generation DNA sequencing target enrichment design comprising a 600 cancer-related gene panel and 100 epigenetics-related genes, comprising the exomes +/- the promoter regions. The target region spanning 7 MB (1x106 base pairs) was designed to cover meaningful portion of genomic, cancer-related sites with a strong emphasis on epigenetic regulators (histone modifiers, chromatin modelers, histone chaperons). Additionally, RNA sequencing was performed to better understand the transcriptomic profiling in the malignant progression of GBM. Targeted sequencing of GBMs demonstrated different genetic drivers (including well known EGFR, TP53, PDGFR and PTEN mutations) and numerous genetic alterations in genes responsible for histone modifications, chromatin remodeling and DNA damage repair. In recurring GBMs we observed only a subset of mutations coming from primary tumors, suggesting a specific path of clonal evolution.
October 18th 2018
A talk given by dr Adam Filip entitled "Prognostic markers for Breast Invasive Carcinoma (BRCA)". Summary: In order to find features that predict breast invasive carcinoma, Monte Carlo Feature Selection and Interdependency Discovery (MCFS-ID) algorithm was applied on the Cancer Genome Atlas (TCGA) dataset containing information obtained from tissues of patients suffering from BRCA. The dataset was composed of information about methylation (HumanMethylation450), mRNA and microRNA expression levels. During the talk, a set of features predicting presence of cancer and/or its subtypes will be presented.
October 10th 2018
A talk given by dr Michał J. Dąbrowski entitled "REST transcription factor ChIP-seq data of glioma tumor patients in the context of their DNA methylation patterns". Summary: The aim of the study is to obtain a better picture of how DNA methylation may affect the binding of REST transcription factor to DNA. This association is studied in the context of discerning glioma patients for PA vs. high grades. We seek for the significant differences that have significant functional effect on specific genes expression. The first preliminary results will be shown as well as plans for the further analysis. I look forward for a fruitful discussion.
September 27th 2018
A talk given by dr Rafał Guzik entitled "IDH in Glioma". Summary: IDH1/2 genes suddenly “Big Banged” cancer genetics in 2008. They are continuously being found mutated in 80 % of grade 2 – 3 glioma and secondary glioblastoma (also in 16 to 75 % tumors of other tissues). Why? What is the mechanism? Let’s review the basics...
September 20th 2018
A talk given by dr Magdalena Mozolewska entitled "Implementation and tests of carbon nanoparticles in coarse-grained UNRES force field.". Summary: Nanoparticles are relatively new, but a rapidly increasing area of interest by scientists. Due to their unique properties, Carbon Nanotubes (CN) and fullerenes can be used in a variety of fields e.g. an industry, military or in a medicine, as carriers of small ligands into target cells e.g. in a cancer therapy, or CNs themselves can be used as ligands, interacting with receptors as their antagonists. Very high computational cost of all-atom force fields limits severely a time-scale of the simulations and a size of the investigated systems. Therefore, in this work, we introduce the implementation of CNs and fullerenes into the physics-based, coarse-grained UNited RESidue (UNRES) force field, which due to the reduction of the number of interaction sites, provides 3-4 order of magnitude speed-up comparing to the all-atom force fields. We simplified representations of CNs and fullerenes to the immobile outer field with adjustable dimensions, which interacts with proteins through Kihara potential. Simulations of protein-nanoparticle systems were tested for the ability to keep constant energy and influence of nanoparticle on temperature stability of the thermostats. Then new potential was utilized to study interactions of various sizes of CNs and fullerenes with different proteins. Our results show that the extended UNRES force field can predict binding of these proteins to CNs correctly and observed the influence of carbon nanoparticles on the protein-structure stability is in a good agreement with the experimental observations.
September 6th 2018
A talk given by Ilona Grabowicz entitled "Towards finding hyperactive TADs in glioma.". Summary: In development of glioma, various genes whose expression becomes altered, play a different roles. Majority of genes are localized within TADs and in this talk the author will present a method for finding hyperactive TADs in glioma and how this method can be applied to other datasets.
June 14th 2018
A talk given by dr inż. Michał Dramiński entitled "dmUtils: a new R helper package to support mcfs analysis.". Summary: During this talk the author of rmcfs package will present practical usage of the new dmUtils package together with new version of rmcfs. The package implements set of functions that simplifies analysis of decision tables as well as more advanced analysis of the top features obtained from MCFS. Some information about upcoming version of rmcfs 1.3 will be also discussed. Entire talk will be more live demo than typical presentation.
May 24th 2018
A talk given by Agata Dziedzic entitled "Overview of DNA methylation analysis - from raw sequencing data to first results.". Summary: We investigated 21 brain tumour samples representing various glioma grades: pilocytic astrocytoma (PA; grade I; n = 7), diffuse astrocytoma (DA; grades II and III; n = 7) and glioblastoma (GB; grade IV; n = 7). We used SeqCap Epi CpGiant Methylation panel and performed bisulphite conversion followed by Illumina sequencing. We obtained at average 23 mln sites per each sample. From those 23 mln sites ~ 3.3 mln were in CpG context and ~ 19.7 mln were in non-CpG context. For further study we only selected sites that were in CpG context and with coverage above 10 reads. We will present first results of differentional methylation calling and methylation patterns across samples form various tumor grade.
May 17th 2018
A talk given by prof. dr hab. Jacek Koronacki entitled "Neo-Darwinism is wrong ...". Summary: Neo-Darwinism is wrong with its emphasis on blind chance as the origin of variation. "Blind chance is necessary, but the origin of functional variation is not at the molecular level" writes Denis Noble in his survey published in Interface Focus 7 (2017), http://dx.doi.org/10.1098/rsfs.2016.0159. He writes also: "Stochasticity is harnessed by organisms to generate functionality. Random-ness does not, therefore, necessarily imply lack of function or ‘blind chance’ at higher levels. [...] The order originates at higher levels, which constrain the components at lower levels. We now know that this includes the genome, which is controlled by patterns of tran-scription factors and various epigenetic and reorganization mechanisms." In my talk, borrowing from Noble's paper, I'll sketch where to look for justification of those claims.
May 10th 2018
A talk given by dr Michał J. Dąbrowski entitled "Unveiling informative DNA methylation site and gene expression profiles that contribute to glioma patients survival". Summary: There is a grooving evidence that molecular markers are important for diagnosis and prognosis of glioma patients. In the Central Nervous System tumor classification provided by World Health Organization in 2016 there is a number of molecular features that were included in the classification together with histology. In this project in order to find clinically useful prognostic markers for glioma patients’ survival, we employed Monte Carlo Feature Selection and Interdependencies Discovery (MCFS-ID) algorithm on DNA methylation (HumanMethylation450 platform) and RNA-seq datasets from The Cancer Genome Atlas (TCGA) for 88 patients observed until death. Using the machine learning methods, we discovered a set of significant features, mainly DNA methylation sites that enable for an accurate distinction between glioma patients with short and long overall survival (OS) i.e. days to death. The significance of the discovered features was confirmed on an independent glioma study cohort. We found that those features are much better predictors of patients’ OS than the previously described molecular markers (such as, e.g., IDH, ATRX, DAXX mutation status). Finally, we found that the most important methylation feature (cg15072976) overlaps with RE1 Silencing Transcription Factor (REST) binding site, is functional and its methylation status affects transcription factor binding in U87glioma cells as evidenced by gel shift assay.
April 26th 2018
A talk given by dr Rafał Guzik entitled "IDH (isocitrate dehydrogenase) in glioma". Summary: IDH1/2 genes suddenly “Big Banged” in cancer genetics in 2008. They have being continuously found to be mutated in about 80 % of grade 2 – 3 gliomas and secondary glioblastoma, in 75 % of cartilaginous tumors, intrahepatic cholangiocarcinoma (23 %), close to 20 % of acute myeloid leukaemia, thyroid carcinomas (16 %), and other tumors. Why? What is the mechanism? Let’s review the basics...
April 19th 2018
A talk given by dr Magdalena Mozolewska entitled "De novo motif search". Summary: De novo motif search is a commonly used method in bioinformatical studies to search and identify specific sequences in a given DNA sequence for further analysis, e.g. an identification of transcription factors motifs. The main hypothesis is that specific transcription factors of the selected genes promotor regions, can distinguish the grade of the glioma tumor. In this study, the DNA sequences were obtained from glioma patients and cover active promotors that are used in the de novo motif prediction to identify transcription factors important in development of glioma.
April 5th 2018
A talk given by Ilona Grabowicz entitled "Investigating microbiome of Down's Syndrome model mice and its changes induced by high-fat diet". Summary: People suffering from Down's syndrome are characterized by malfunctioning of the gastrointestinal tract and are more prone to obesity than healthy ones. Recent advances in sequencing technology made it possible to study microbiomes. In my talk I will present results of the analysis of microbiomes of the Down's syndrome mouse models, which are also subjected to high-fat diet, which promotes obesity and has addictive properties. I will show which bacterial species are differentially abundant between these two mouse strains and also diets. I will show the pros and cons of sequencing microbial DNA and RNA. Moreover I will show the results and pitfalls of analyzing the microbial gene expression.
March 22nd 2018
A talk given by dr Michał Dramiński entitled "Analysis of IDH and gene expression levels of Symfonia patients". Summary: This talk is about selection of the most significant genes for IDH status, tumor grade and survival for set of TCGA and Symfonia patients.
March 15th 2018
A talk given by Marta Jardanowska entitled "Motif analysis: Detection of transcription factor motifs within the grade specific active promoters marked by histone methylation". Summary: The talk is about the latest research related to detection of motifs of transcription factors in a set of DNA sequences marked with specific histone methylation in the promotor region of transcribed genes. The first set of methylated promoters comes from the following selection: promotors active in grade one and non-active in the higher grades. The other set is opposite to the first one. There are active promoters from higher grades and non-active in the grade one. The aim of this presentation is to describe biological background of the research and methods used. Some preliminary results will be presented as well as the encounter difficulties and further steps of this analysis.
February 22nd 2018
A talk given by prof. Jacek Koronacki entitled "Issues in Big Data Analysis: Dealing with Data Streams". Summary: After a short intro explaining what is usually meant under the term "Big Data", I shall confine myself to dealing with data streams. I shall report on: (i) current approaches to building classification trees for stationary series; (ii) a simple way to adjust an ensemble size in stream data classification; and (iii) a not so simple way to develop regression models for nonstationary data streams. My emphasis will be on methods with sound mathematical foundation (albeit with all the proofs and mathematical details skipped). The talk will be based on the work of Prof. Leszek Rutkowski and his research group at Częstochowa University of Technology.
February 1st 2018
A talk given by dr Paweł Teisseyre entitled "Introduction to multi-label classification". Summary: Recently multi-label classification has attracted a significant attention in machine learning community. In standard classification problem we investigate the relationship between response variable (e.g. occurrence of cancer) and features (e.g. expressions of genes, SNPs or methylations). In multi-label problem we consider many response variables (labels) simultaneously (e.g. different diseases). The goal is to build a model which predicts response variables (labels) based on feature values. Multi-label methods has been successfully applied in many domains, an interesting example is modeling mulimorbidity (co-occurence of many diseases in one patient). In my talk I will introduce the basic methods: Binary Relevance, Label Powerset and Classifier Chains and discuss the available software.
January 18th 2018
A talk given by dr Bartosz Wojtaś entitled "Next generation sequencing focused on transcriptomics". Summary: Next generation sequencing technology has emerged as a leading biomedical technology of XXI century. It has revolutionized modern science and medicine. In my talk I want to explain basics of that technology for non-expert audience and discuss in details potential of transcriptomics analysis. In my talk I will also discuss differences between two most commonly used global transcriptomics analysis - microarrays and RNAseq.
January 4th 2018
A talk given by Marta Jardanowska entitled "Network activity of selected neurons of the C. elegans nematode in interaction processess of the organism with its environment". Summary: Caenorhabditis elegans is a popular model organism in studies of embryogenesis, morphogenesis, development and neurobiology. Because of the good knowledge of anatomy of this nematode, understanding its neural networks can provide many valuable solutions to today's problems referring to the functioning of nervous systems and the functioning of the human brain. Because C. elegans is a simple organism, therefore, in the current state of the knowledge, leading both experimental and computational research, we have a real possibility of obtaining reliable and accurate results.
Seminars 2017
December 21st 2017
A talk given by Ilona Grabowicz entitled "Investigating brain gene expression and gut microbiome changes and its interactions with the host upon diet-induced behaviour changes". Summary: Diet is influencing organisms in numerous ways. In the seminar, it will be described how 'cafeteria diet' can affect behavior of mice and gene expression in a TAD-dependent manner in different parts of the brain. Moreover, it will be shortly discussed how the high-fat diet and differences in mice strains can affect also behavior of mice and also composition of their gut microbiota.
December 7th 2017
A talk given by Agata Dziedzic entitled "Fungal lifestyle reflected in serine protease repertoire". Summary: Fungi are able to switch between different lifestyles in order to adapt to environmental changes. Their ecological strategy is connected to their secretome as fungi obtain nutrients by secreting hydrolytic enzymes to their surrounding and acquiring the digested molecules. We focus on fungal serine proteases (SPs), the phylogenetic distribution of which is barely described so far. In order to collect a complete set of fungal proteases, we searched over 600 fungal proteomes. Obtained results suggest that serine proteases are more ubiquitous than expected. Our study present a comprehensive evolutionary history of fungal SP families in the context of fungal ecology and fungal tree of life.
November 9th 2017
A talk given by dr Michał J. Dąbrowski entitled "Methylation patterns in cancer patients". Summary: DNA methylation is one of the most frequently studied epigenetic elements in mammals. It regulates gene expression in cells – activation and silencing depending on the cell function specificity. It is also associated with histone modifications and by that with the architecture of a chromatin. The covalent addition of a methyl group is most frequently present in cystosine within CpG dinucleotides. The concentration of (CpG)n is called CpG islands. The majority (~70%) of human promoters located near transcription start site (TSS) of a gene contain a CpG island. During an organism development many epigenetic factors may change, including CpG methylation and chromatin reorganization. DNA methylation is known to be abnormal in all studied forms of cancer. This talk is focused on how this occurs, what is its role in tumorigenesis and how is it used for cancer classification purposes and prediction of patients prognosis.
October 26th 2017
A talk given by dr Magdalena Mozolewska entitled "Molecular modeling of protein systems". Summary: The Iron-Sulfur Clusters (ISCs) are highly spread among the organisms from prokaryote to eukaryote. They are the crucial part of many cellular processes like electron transfer, redox reactions, and catalysis of chemical reactions. Many proteins require presence of chaperons to achieve and stabilize their structures. In yeast Saccharomyces cerevisiae the Fe/S cluster biogenesis cycle is one of the mechanisms, which are necessary for the organism, but despite the numerous studies, still not well known. To investigate such intricate system coarse-grained UNRES force field was used to obtain the structure and dynamics of the Isu1-Jac1 and Isu1-Jac1-Ssq1 complexes.
October 12th 2017
A talk given by dr Michał Dramiński entitled "Monte Carlo Feature Selection and Interdependencies Discovery (MCFS-ID). Theory and practical usage of the algorithm". Summary: The aim of the lecture was to present and teach an innovative approach (MCFS-ID) to select informative features from high dimensional complex Life Science data sets using R. R is powerful environment and language for data processing and it supports various machine learning, statistical, optimization and data mining functions. The package rmcfs is not only MCFS-ID implementation for R – it is a complex solution for feature selection, discovery of feature interdependencies (ID) as well as evaluation and later processing of the obtained result.