Welcome to the “Coffee break with TRANSFAC”
A series of online sessions hosted by Dr. Alexander Kel, CEO geneXplain GmbHThe next Coffee break with TRANSFAC will be held on March 19th at 10 AM CET
Fill in this form to receive the event joining link
MATCH Suite software demo
In this webinar Dr. Alexander Kel, the CEO and CSO of geneXplain GmbH, together with Daria Stelmashenko, the lead of the MATCH Suite development team, will demonstrate the abilities of the MATCH Suite component of TRANSFAC 2.0.
The following main questions will be in the focus of this seminar:
– What is MATCH Suite and in which cases it should be used?
– What are the analysis pipelines underlying the automatised MATCH Suite workflows?
– How is the analysis launch performed in MATCH Suite? How to interpret the obtained results and where to use them?
Both: single gene analysis and gene set analysis will be performed, and the obtained results will be discussed, including the interactive user-friendly interface allowing you to fine-tune your results by applying various filters, as well as the automatically generated analysis report.
About the Coffee breaks with TRANSFAC
This initiative of Q&A sessions with a leading bioinformatics expert Dr. Alexander Kel is intended to support all researchers out there that are interested in the area of applied bioinformatics.
Come to our sessions as a simple listener, or become an active participant of this recurrent event and ask Dr. Kel your own questions in order to emphasise the direction of the live discussion.
You can send your questions via the form below or by email: [email protected] with a subject “Question to Dr. Kel”.
Questions can also be asked live during the online event.
Ask your question by filling in this form:
What to ask about?
Any question that you need assistance with while performing your bioinformatics analysis, e.g.
- Promoter analysis? Pathway analysis?
- Can AI analyze NGS data?
- How to combine DNA methylation and metabolome data?
- What to start from and how to interpret the obtained results?
and much more….
Check out the video records of the previous “Coffee break with TRANSFAC” sessions
- Promoter analysis of model organisms (part 2)
- Promoter analysis of model organisms (part 1)
- How to find tissue-specific TF target genes
- Your first TRANSFAC analysis
- How to find TRANSFAC motifs in ATAC-seq peaks? (part 2)
- How to find TRANSFAC motifs in ATAC-seq peaks? (part 1)
- TRANSFAC pathways. How to find them?
- What is the way from TFs to pathways?
- How to build gene regulatory networks using TRANSFAC, Jupyter Notebook and API
- Master-regulators of Glioblastoma pathways – special session dedicated to the World Brain Tumor Day
- Promoter analysis of model organisms
- How to apply TRANSFAC for analysis of cancer mutations
- How to find motifs created or destroyed by SNPs
- De novo motifs: how to find and use them
- How to compute transcription factor binding affinity
- How to find a list of tissue-specific TF target genes
- Promoter analysis? Pathway analysis? What to start from? How to interpret the results?
12 March 2024, the seventeenth “Coffee break with TRANSFAC” session
Promoter analysis of model organisms (part 2) Click here to view the event details
Your questions addressed within this session:
01:11 Number of different positional weight matrices (PWMs) or motifs contained in TRANSFAC for various organisms
02:56 Matrix construction for TRANSFAC database
04:08 Site enrichment analysis using TRANSFAC PWMs
06:07 Composite elements of transcription factors in gene regulation
07:16 Genetic algorithm for identification of TFBS combinations
10:45 geneXplain platform overview
13:03 Example: RNA-seq and ChIP-seq data analysis from Drosophila melanogaster (GEO: GSE149116) Spoiler: human ZKSCAN3 and Drosophila M1BP are functionally homologous transcription factors acting as master transcription factors autophagy regulation! It appeared that transgenically expressed human ZKSCAN3 gene in drosophila started to regulate many drosophila genes binding to the same sites as endogenous M1BP factor.
18:20 Starting an integrated ChIP-seq and RNA-seq analysis inside the geneXplain platform: using DEGs from transgenic drosophila when compared to the wild type, and ChIP-seq peaks for the antibody for human ZKSCAN3 transcription factor from GEO: GSE149116
24:17 Is there a “proper” promoter length for performing the analysis?
28:35 What is the advantage of using the geneXplain platform to see the results of this type of experiment (integrated ChIP-seq and RNA-seq analysis) instead of using the IGV like apps?
32:39 Analysis of DEGs regulation in the geneXplain platform: filtering of a table with DEGs and selecting the ChIP-seq peaks located near the significantly differentially expressed genes using the filter one track by another function
40:38 Finding enriched motifs in tracks constructed as ChIP-seq peaks located around the upregulated genes
45:47 Results overview and interpretation for the identification of enriched motifs in tracks (site search summary and MATCH track)
50:35 Launching the CMA (Composite Module Analyst) analysis, and CMA results overview
57:53 What can TRANSFAC do in relation to enhancer sequences?
01:03:36 Are there also curated СhIP-seq datasets in TRANSFAC that can be used / compared along with own data?
01:07:33 Experimentally proven TFBS in TRANSFAC
01:09:24 The genome used for human is hg38, but is it the canonical version or the whole assembly?
4 March 2024, the sixteenth “Coffee break with TRANSFAC” session
Promoter analysis of model organisms (part 1) Click here to view the event details
Your questions addressed within this session:
02:37 Model organisms in TRANSFAC database
06:37 geneXplain platform overview: model organisms available in the system by default
09:20 viewing TRANSFAC motifs in the geneXplain platform interface; what is a profile and how to select it
12:08 Construction of a PWM (positional weight matrix). Cutoff and core cutoff.
17:13 Mini pig model for atherosclerosis study
19:03 Importing data to the geneXplain platform via ftp
21:35 Principles of site enrichment analysis using the TRANSFAC database
23:59 Upload of a new genome to the geneXplain platform (from NCBI)
27:33 Genome browser visualization of the newly uploaded genome
32:47 Identification of DEGs (differentially expressed genes) with Limma; filtering of the obtained results
39:34 Extraction of promoters of the pig genes using the “process track with sites” function of the geneXplain platform and the coordinates of pig genes taken from the MART database (EBI)
43:23 Search for enriched TFBS using the Site search on gene set function of the geneXplain platform
49:40 Finding combinations of motifs belonging to TFs working together using the CMA (Composite Module Analyst)
55:05 Upstream analysis with feedback loops – results overview
01:02:26 Can you start the analysis from differentially expressed genes (Limma output)?
01:07:42 Converting between different accession numbers in the geneXplain platform
01:11:02 Overview of the integrated HumanPSD+TRANSPATH+TRANSFAC database interface
6 February 2024, the fifteenth “Coffee break with TRANSFAC” session
How to find tissue-specific TF target genes Click here to view the event details
Your questions addressed within this session:
0:44 How to find all target genes of NFkB transcription factor? Clue: in which tissue?
02:17 MEF-2A transcription factor and its expression profile
04:20 Which TFs are regulating my genes specifically expressed in muscle tissue
05:42 MATCH Suite gene set analysis launch
09:41 MATCH Suite gene set analysis results overview
22:57 MATCH Suite gene set analysis report overview
26:31 Tissue-specific gene regulation
28:30 TFClass transcription factors classification
30:50 Expression patterns of transcription factors coming from one family
33:55 Constructing tissue specific gene set in MATCH Suite
36:26 Gene list optimization by GO terms in MATCH Suite
41:39 Heatmap of GO terms and TF motifs produced by MATCH Suite
44:55 MATCH Suite results overview from the geneXplain platform interface
48:41 In my research area (plant pathogenesis) TFs are not enough available in Jaspar. Can I proceed with such species when experimentally validated TFs are not available in the database?
52:31 We found an unknown Zinc finger protein. Can I use your tool to predict the target genes?
55:20 Any cross-reference data base on transcription factors?
57:29 If we are analyzing a TF gene and the question is to find the downstream genes of this TF in genome of that organism. How the downstream genes can be identified? (check out this other video demonstrating how all TFBS of one transcription factor can be found)
23 January 2024, the fourteenth “Coffee break with TRANSFAC” session
Your first TRANSFAC analysis Click here to view the event details
Your questions addressed within this session:
02:32 Gene list analysis in the geneXplain platform (search for TFBS in promoters of genes in focus)
04:48 Upload of table with differentially expressed genes (DEGs) to the geneXplain platform
07:55 Saving a single gene from a gene list for further single gene promoter analysis
08:33 Prediction of transcription factor binding sites (TFBS) in the promoter of a single gene (site search on gene set with 1 gene in the list)
14:30 Visualization of found sites in the promoter model of the studied gene
16:25 Opening the track of found sites in the full genome browser visualization
17:37 Dragging and dropping genes, repeats, and variations tracks to the genome browser visualization
19:19 Export of the obtained results: saving the table of found sites and export of found sites as a BED file
22:29 – What do you mean by property score and core score? Does it give p-values of the predicted transcription factors? On what basis can we select the TFs?
23:41 The model of eucaryotic gene regulation
28:37 How are the matrices built? Additive model of sites prediction
30:41 From count matrix to the frequency (probability) matrix. Log-odds score of sites (used in HOMER)
34:08 MATCH site score: multiplication by the information vector
36:01 The score cutoffs selection (which cutoffs to select for your MATCH analysis)
37:44 The profile used in MATCH analysis: collection of matrices and their cutoffs
40:38 Comparison of MATCH algorithm with log-odds or additive algorithm
41:31 Protein-protein interactions compensating the actual binding energy of individual site. Why cutoffs should not be fixed at all and can be automatically found by the algorithm
44:38 Working with gene sets describing the studied condition, e.g. DEGs and non-changed genes that didn’t respond to the condition (yes and no sequences)
46:06 Automatic cutoffs identification that maximize the difference between the sites found in yes and no sets
47:15 Site search on gene set using a No set (non-changed genes)
47:55 Profiles used for site search
51:51 Is it possible to search for motifs for several TFs taking into account their interaction?
53:17 Overview of the output table of gene set analysis with No set and cutoffs optimization
54:53 Visualization of found sites in the models of promoters of the genes from the input gene set
59:26 Is the promoter browsing interval fixed? (1000 bp upstream and 100 bp downstream the TSS?)
01:04:18 Is it possible to find TFs if we don’t have any ChIP-seq data for the particular gene?
01:07:46 How do you backtrack a particular TF from a matrix predicted with your tools? How do I know which protein is actually regulating my gene? What are the filters I can apply to all TFs corresponding to that matrix in order to select the most probable one for my case?
01:12:56 On the basis of only the score, how can you select the TFs, as TRANSFAC gives a huge list of TFs, how to filter it?
Your questions addressed within this session:
00:56 ATAC-seq data: what is it and how is it processed?
05:05 The model of intracellular gene regulation; master regulators
07:17 Algorithms for ATAC-seq data analysis
08:05 Desmoplastic small-round-cell tumor: GSE226670 dataset (ATAC-seq + RNA-seq)
10:55 SRA, table data, and BED files upload to the Genome Enhancer
18:19 Annotation diagram construction in Genome Enhancer
20:48 Analysis launch in Genome Enhancer: disease selection and specification of conditions for comparison
22:47 Genome Enhancer results: report overview
29:50 Transcription factor binding sites in ATAC-seq peaks
36:45 Search for master regulators; positive and negative feedback loops
42:21 Selection of prospective drug targets and associated treatments
44:30 Genome Enhancer results visualization in the geneXplain platform interface
52:41 How can quantitative information be applied in ATAC-seq peaks analysis? Interpretation of the ATAC-seq peaks height: how often a particular location is accessible for the enzyme? Selection of motifs appearing more frequently in high peaks and less frequently in low peaks. Alternative quantitative information: expression level of a gene.
Your questions addressed within this session:
00:37 ATAC-seq data – what is it?
04:18 How ATAC-seq data is analyzed
06:50 The general approach towards upstream analysis
11:23 Algorithms for ATAC-seq data analysis, peak calling approaches
13:53 GSE226670 dataset on acute myeloid leukemia cell line: integrated RNA-seq and ATAC-seq analysis devoted to BPTF knock-out compared to wild type
18:13 Analysis of GSE226670 ATAC-seq and RNA-seq data using the geneXplain platform
19:17 SRA data upload to the geneXplain platform using the SRA ID
21:21 FASTQ files mapping to the genome in the geneXplain platform (with Bowtie2)
22:59 BAM file visualization on genome browser in the geneXplain platform
24:24 Peak calling with MACS2 from a BAM file in the geneXplain platform
27:43 Search for transcription factor binding sites (TFBS); search for enriched TFBS in ATAC-seq peaks areas
29:42 Uniting all peaks from one experiment (join tracks function of the geneXplain platform)
31:33 Finding peaks from the wild type cells that are not present in the knock-out experiment (intersect tracks function of the geneXplain platform)
32:41 Epigenomics data analysis in the geneXplain platform: site search with TRANSFAC
36:06 Site search results visualization and interpretation
44:00 Genome Enhancer application for integrated ATAC-seq and RNA-seq data analysis. Analysis launch
48:56 Genome Enhancer analysis report. Enriched transcription factors and key master regulators. Intersection of peaks from the wild type cells with the upregulated genes
54:57 The YY1 (Yin Yang) binding sites balance; the Yin Yang factor with controversial regulatory functions
58:40 Further steps after TFBS analysis: master regulators in Genome Enhancer report and prospective drugs overview
Your questions addressed within this session:
02:53 The basic model of gene functioning and regulation
05:24 Brief overview of transcription factor binding sits (TFBS) prediction
09:06 How to find master regulators that control the activity of identified transcription factors
14:18 Walking pathways: pathway rewiring
19:27 GeneXplain platform brief interface overview
20:52 Example of gene expression analysis data from GSE66789 (normalized counts – FPKM – from osteosarcoma cell lines with induced Myc gene) + proteomics data (logFC of protein expression in Myc induced cells compared to controls)
23:25 Data upload to the geneXplain platform
25:47 Finding differentially expressed genes (DEGs) in the geneXplain platform interface
27:11 Filtering of DEGs, selection of upregulated genes
28:31 For RNA-seq data analysis can TPM be used as well? (instead of FPKM)
29:26 Construction of the control set of genes (needed for search of enriched motifs in promoters (TFBS search)
31:11 Finding binding motifs
35:21 Overview of the site search tabular results; Yes-No ratio
38:18 Genome Browser visualization of the found motifs
43:35 Upstream analysis: searching upstream of the identified transcription factors. Identification of master regulators.
53:49 The constructed network is generated by the means of the literature search or is it from the specific experiment?
55:35 Visualization of TRANSPATH pathways in Cytoscape
59:20 Different weights of the network proteins based on the betweenness centrality
Your questions addressed within this session:
01:57 What is the way from transcription factors to pathways? How do transcription factors operate?
03:15 What happens after we have identified transcription factor binding sites (TFBS)?
06:33 Prediction of TFBS in promoters of genes using positional weight matrices (PWMs)
08:20 Combinations of transcription factors, Composite Module Analyst (CMA)
10:48 Search for master regulators
21:15 Walking pathways
31:22 Loading of RNA-seq data (normalized counts) to the geneXplain platform
34:50 Filtering of genes; selection of up and down regulated genes and non-changed genes
36:11 Identification of enriched motifs in promoters of genes using TRANSFAC
41:06 What are the transcription factors regulating your genes (after finding the enriched motifs)
44:33 Going upstream from the identified transcription factors
45:01 Regulator search: how to find master regulators
50:10 Network visualization of master regulators
51:03 Annotating the visualization diagram with expression data
51:54 Applying different layouts to the constructed network
52:55 Opening the TRANSPATH network in Cytoscape
55:45 Can own interaction data be introduced by the user? REACTOME, Recon, GeneWays
58:20 Can you mix different databases? Adding a file with different interactions from various databases
01:01:32 Is it possible to construct the master regulator-TFs network without proteomics context? Are there any reconstruction examples of this type network for microorganisms?
01:03:53 How can we plot the signaling pathway for tuberculosis host hub genes (or any genes) using the geneXplain platform?
01:09:06 Is it possible to discover new regulatory pathways based on expression data?
Your questions addressed within this session:
00:20 Gene regulatory networks – what are they and how to construct them
21:12 MATCH Suite and gene regulatory networks based on predictions of transcription factor binding sites
27:35 MATCH Suite results visualization: constructing gene regulatory networks using Python* in Jupiter notebook (geneXplain platform API)
*The Python code shown in the online demo is available here
40:42 From which online software I can get the sequences of transcription factors?
50:30 Can TRANSFAC be applied for other species, e.g. drosophila or plants? What are the limitations?
Your questions addressed within this session:
00:02 Introduction: TRANSFAC in Cancer research
00:56 World Brain Tumor day
01:45 Transcription factors: how do they regulate their target genes
02:21 Positional weight matrices – TFBS model
03:07 Searching for TFBS enrichment
03:48 Composite complexes of transcription factors
04:47 Search for Master Regulators
06:26 Intracellular signal transduction: complex cascades consisting of signaling reactions
08:18 Construction of potential network of possible reactions within the cell
09:01 Modeling of regulation of a particular set of genes
11:59 Algorithm of Master Regulators searc
13:42 Integration of context protein expression information in the search for Master Regulators
15:23 Verification of Master Regulators quality
18:08 Walking pathways concept
20:35 How to select the “true” Master Regulator
22:08 Therapeutic targets and biomarkers of the studied processes
22:59 Methylation marks located in regulatory regions of Master Regulator genes are good diagnostic or prognostic biomarkers
23:18 The Upstream Analysis concept (integrated promoter and pathway analysis)
23:58 Genome Enhancer – the fully automated pipeline for prospective drug target identification
26:43 Analysis of Glioblastoma short-term survival patients vs. long-term survival patients using Genome Enhancer (this analysis resulted in the following publication: IGFBP2 Is a Potential Master Regulator Driving the Dysregulated Gene Network Responsible for Short Survival in Glioblastoma Multiforme)
41:47 Switch from Genome Enhancer to the geneXplain platform interface with extended functionality. Mapping of prospective Master Regulators to diseases in which they are known biomarkers.
47:38 Immunotherapy sensitivity prediction for Glioblastoma patients
51:26 Are there transcription factor and enhancer databases for nonhuman primates?
56:30 Is there a classical publication involving geneXplain related to the topic?
Yes, please check those on Glioblastoma: (1) and (2), this one on the Walking Pathways concept, and other publications that can be found here.
58:57 Can we check these results with the data on mutations in samples? Won’t this help find up- or down-regulated genes with respect to their mutation profile?
01:04:53 If I have RNA-seq data for my patient, which additional omics data could improve the analysis with your tool effectively? E.g. variants, protein expression, methylation?
Your questions addressed within this session:
02:47 TRANSFAC brief overview
05:08 Site enrichment analysis
07:59 Site combinations: complexes of transcription factor binding sites; composite modules
14:27 TRANSFAC application for different model organisms
21:27 Example on drosophila genes analysis from GSE149116 – combination of RNA-seq and ChIP-seq data
22:31 What are the target genes of my transcription factor? Is such question correct?
23:16 Human TF was put as a construct and expressed in Drosophila –> Human TF started regulating the drosophila genes! Wow!
25:29 Different experiment types being united in one analysis: intersection of RNA-seq data with ChIP-seq data
28:25 Drosophila data analysis in the geneXplain platform (on the example of GSE149116 dataset)
41:02 How site search summary looks like in the geneXplain platform
42:21 Identification of “master” transcription factors working together in combinations; CMA (Composite Module Analyst) analysis in the geneXplain platform
46:44 Can we use other genomes (other than human) for TFBS analysis with TRANSFAC in the geneXplain platform
47:15 How to upload the palm genome to the geneXplain platform
50:33 Extraction of promoters of genes responsible for palm resistance to different bacteria or infections. “Master” transcription factors regulating the resistant genes in the palm.
54:35 What are the “standard” genomes in the geneXplain platform
56:01 How can I find master transcription factors that regulate specific pathways in Arabidopsis
58:17 Which profile (collection of PWMs – positional weight matrices) to use for the analysis?
Your questions addressed within this session:
- Mutations in cancer: how to analyze them using TRANSFAC?
- Introduction to Genome Enhancer – an automatized pipeline for multi-omics data analysis. Analysis of cancer mutations using Genome Enhancer.
- Lung cancer cell lines analysis using Genome Enhancer. A more detailed video devoted to this example can be found here.
- By which authority is Genome Enhancer certified for application to patient data in hospital?
- Mutations located in non-coding regions of genes: can they destroy or create TFBS?
- Can we predict the effect of these gene mutations on drug binding to targets?
- Analysis of binding sites compositions: search for co-factors – transcription factor binding sites (TFBS) working together. Identification of TFBS compositions around clusters of mutations when compared to regions without mutations.
- Switch to the geneXplain platform perspective: a more detailed view on results produced by Genome Enhancer pipeline in regards to the TFBS analysis.
- Brief overview of the next sections of Genome Enhancer report: pathway analysis, identification of master-regulators in networks and selection of prospective drug targets and associated treatments.
Your questions addressed within this session:
- SNP data analysis: motifs destroyed and created by SNPs – how to find them and what is their effect?
- Overview of the geneXplain platform interface
- Analyzing SNP data with the geneXplain platform (Variant Analysis)
- Is it possible to use TRANSFAC to calculate p-values for binding with a specific allele at the SNP and to rank predicted TFs based on their binding score?
Your questions addressed within this session:
- De novo motifs. What are they? How to find and use them? (short intro)
- Why are there so many motifs (PWMs – positional weight matrices) collected for 1 transcription factor? All of them are different. Which one is “true”? (spoiler: they ALL are!)
- Searching for de novo motifs: when and why do we need new motifs? How are they discovered?
- ChIPMunk motif discovery tool – what is it?
- De novo motif search using the geneXplain platform
- Can I perform TFBS search with TRANSFAC alone?
- Can geneXplain platform find mutation in the intron (when compared with reference sequence, e.g. hg38)?
Your questions addressed within this session:
- How to compute transcription factor binding affinity?
- What is the difference between the 5’ prime and the 3’ prime promoters?
- I have two genes with overlapping promoters, but different TFBS, can you comment on this?
- If I design a minimal promoter, what would be the basis for screening and selecting the best TFBS to include in my minimal promoter if I use the MATCH Suite tool? In brief, how to prioritize these TFs to select the most relevant of them?
- Could you explain what is the difference between p-values in the MATCH Suite analysis report?
- For one TF there can be numerous PWMs defined by a number in the name such as V$AP1_02. Does this number correspond to a new version of the PWM based on the alignment of known TFBS? Do you advise to use the last PWM built on the alignment of known binding sites as the most updated one or do we need to use all PWMs for a TF to perform the prediction for a particular TF?
- Alexander, as I understand, the affinity score(s) is a floating number. How does the program define the cut-off (maybe I missed that) – every time depending on the experiment, or it is set up on an average basis?
- Is it important to have cell specific analysis and not only tissue or organ specific one?
Your questions addressed within this session:
- TFBS prediction in DNA based on positional weight matrices
- How to define the cutoffs for binding sites scores
- In calculating FP and FN you need to know the true binding sites. How are they determined?
- I have some questions about my data derived from the TRANSFAC 2012 professional matrices. 1) Some of the genes of interest have binding sites for IRF (without further specification of the subtype), STAT (again, without further specification of the subtype). Can you explain what this means? 2) Some other genes have binding sites for STAT1STAT1 – does this mean STAT1 dimer, also known as the GAS element?
- I am interested to find out what TFs, promoter & enhancer regions, and epigenetic signatures are relevant for the regulation of specific genes in rat hippocampal pyramidal and inhibitory neurons. I know this is not an easy task, and needs not only a bioinformatic but also an experimental approach. I am not aware how transcriptome (RNAseq) and epigenetic (ATAC-seq) data are integrated in the TRANSFAC platform and whether TRANSFAC could help in refining this.
Your questions addressed within this session:
- How can I start a TRANSFAC bioinformatics analysis? What types of data do I need?
- What is the significance of a transcription factor binding site on plus or minus strand of the gene?
- New transcription factors – do you survey those all the time? How many targets are necessary for inclusion of one (creation of a matrix) in the program?
- What advantages does TRANSFAC have over the HOCOMOCO and JASPAR databases?
- Regarding TF binding analysis: even after we get a putative motif for TFs, how many TF could possibly bind at a single site on the genome, depending on a significance cut-off, I find TF binding sites from 2-3 to almost 40-50 at the same location. How do we filter the noise or get accurate results?
- Is it possible to use directly gene sequences to search TFBS? For example sequences from Ensembl database? FASTA file?
- Either on + strand, or on – strand, one should read the binding site from 5′ end to 3′ end, is it correct?
- Can a gene have two active promoters and corresponding TSS in the same cell?
- In order to investigate how much a genetic mutation could affect a TF binding site, is it possible that such a mutation could affect multiple TFs binding at that region? Would it be advisable to focus only on TFs based on their binding score?
- The co-occurrence of TFs proposed by TRANSFAC is based on the knowledge coming from different experiments, so what are the chances of co-occurrence actually?