1-BIN-301, 2-AIN-501 Methods in Bioinformatics, 2023/24

Introduction · Rules · Tasks and dates · Materials · Moodle
Quizzes can be found in Moodle.
Homework assignments and journal club papers can be found in Tasks and dates.
Exam rules, example questions and syllabus
Groups for journal club have each their own group in Moodle.

CI-en-db: Rozdiel medzi revíziami

Prejsť na: navigácia, hľadanie
(Gene expression data)
(Sequence motifs, program MEME)
Riadok 152: Riadok 152:
===Sequence motifs, program MEME===
===Sequence motifs, program MEME===
* Vazobne miesta transkripcnych faktorov sa casto reprezentuju ako sekvencne motivy
* Program MEME gets a group of sequences and find a motif they have in common
* Ak mame skupinu sekvencii, mozeme hladat motiv, ktory maju spolocny
* Based of EM algorithm and probabilistic models
* Znamy program na tento problem je MEME
* Ak mame skupinu sekvencii, mozeme hladat motiv, ktory maju  
* Chodte na stranku http://meme.nbcr.net/
* Go to http://meme.nbcr.net/ select MEME tool
* Zvolte nastroj MEME a do okienka "actual sequences" zadajte [http://compbio.fmph.uniba.sk/vyuka/mbi-data/cb11/seq.fa tieto sekvencie]
* As "actual sequences" enter [http://compbio.fmph.uniba.sk/vyuka/mbi-data/cb11/seq.fa this data]
* Pozrite si ostatne nastavenia. Co asi robia?
* If the server computes too long, you can see precomputed results [http://nbcr-222.ucsd.edu/opal-jobs/appMEME_4.9.114170270799951339152135/meme.html here]
* Ak server pocita dlho, mozete si pozriet vysledky [http://nbcr-222.ucsd.edu/opal-jobs/appMEME_4.9.114170270799951339152135/meme.html tu]
===Genes and proteins===
===Genes and proteins===

Verzia zo dňa a času 21:13, 25. marec 2015

This page is under construction, it contents will change

Introduction to bioinformatics databases and on-line tools

The goal of this excercise is to

  • see results of bioinformatics research in the form of on-line tools used by many biologists
  • get to know some basic tools in case you might want to try your algorithms on biology data
  • review some of the topics from the lectures

NCBI, Genbank, Pubmed, blast

  • National Center for Biotechnology Information http://www.ncbi.nlm.nih.gov/
  • Collects publicly available data in molecular biology
  • We can search for keywords in various databases
  • BLAST finds alignments of query sequence and a specified sequence database
    • convenient, because no need to download large database, but also very slow
  • Try sequence below at http://blast.ncbi.nlm.nih.gov/Blast.cgi
    • the sequence is from the human genome but we will try to find its homolog in chicken
    • choose nucleotide blast, database reference genomic sequence, organism chicken (taxid:9031), program blastn)
    • on which chromosome is the best chicken homolog, what is alignment length, score, E-value, identity level?

UCSC genome browser

  • http://genome.ucsc.edu/
  • nice interface for browsing genomes, lot of data for some genomes (particularly human), but not all sequenced genomes represented
  • also allows custom queries and data download


  • on the front page choose Genomes in the top blue menu bar
  • select genome and its version, optionally enter position or keyword, press submit
  • on the browser screen top image shows chromosome map, selected region in red
  • below a view of selected region and various track with information pertaining to these regions
  • tracks can be switched on and off and configured in the bottom part of the page
    • different display levels, full contains all information but takes a lot of vertical space
  • navigation at the top (move, zoom, etc.)
  • various actions in the menu
  • clicking at the browser figure allows you to get more information


  • Instead of BLAST, UCSC genome browser uses faster but less sensitive BLAT (good for the same or very closely related species)
  • Go to http//genome.ucsc.edu/, choose Blat in the top blue menu bar, enter DNA sequence above, search in the human genome
    • What is the identity level for the top found match? What is its span in the genome? (Notice that other matches are much shorter)
    • Using Details link in the left column you can see the alignment itself, Browser link takes you to the browser at the matching region
  • Go to the browser, switch on Vertebrate net/chain on full
    • this track allows you to move to corresponding parts of other genomes
    • in the chicken chain notice chromosome number of the corresponding region in chicken
  • Optionally, you can try to use BLAT to map the query to the chicken genome directly
    • on the blue bar press genomes, choose vertebrate and chicken, then blat on the top bar in submenu Tools
    • what is not idemtity level and span of the best match? Is it on the same chromosome? How does it compare with the values obtained at NCBI?

Sequencing and assembly

  • UCSC genome browser has numbered version of individual genomes - errors and missing parts are fixed over time
  • Go to genome.ucsc.edu, choose Genomes in the Blue bar, select human, see when were the last version of the human genome added
    • if you are interested in detail, each assembly has a description at the bottom of the page
  • Go to the browser for human assembly hg19, region chr2:110,000,000-110,300,000, you can use this link: [1]
  • Display tracks "Assembly" and "Gap" in the full mode.
    • What is the length of the unsequenced gap in the middle? (you can click on the gap to get details; only an estimate, not sequenced in this assembly)
    • This gap is closed in the most recent assembly hg38. You can have a look by transfering to corresponding region in hg38 - click on the blue bar View -> In other genomes (convert), seelect hg38. Notice that the length of the region shrank from 300,000 to 158,880. So the gap length estimate was not very accurate.

Comparative genomics

Background: HAR1 gene

  • Pollard KS, Salama SR, Lambert N, et al. (2006). "An RNA gene expressed during cortical development evolved rapidly in humans". Nature 443 (7108): 167–72. doi:10.1038/nature05113. pdf
  • Authors found regions with many human-specific mutations but conserved in other mammals (using probabilistic models)
  • 49 statistically significant regions
  • The most significant is HAR1: length 118, 18 substitutions in human, expected value 0.27. Only 2 substitutions between chimpanzee and chicken.
  • Overlaps RNA gene HAR1 (multiple forms)
  • One of the forms is expessed in embryonic neocortex and other parts of the brain

HAR1 and comparative genomics in the browser

  • You can see this region in the browser: chr20:61,733,466-61,733,626 (hg19)
  • Make sure Conservation track is switched on full mode (perhaps press default tracks button)
  • If you zoom in closer, you will see a multiple sequence alignment, with many changes specific to human
  • If you zoom out to a wider region, e.g. chr20:61,733,305-61,733,787, you can look at PhyloP substrack which shows for every base its conservation level - increase conservation over mammals in general in the HAR1 region

Population genomics in the browser

Population genomics studies differences between individuals within species, e.g. between different people

  • Go to region chr2:174,862-436,468 in hg19
  • In section Phenotype and Disease Associations set GAD view track to full
    • This track shows knows associations of particular genetic regions or mutations to diseases
    • You can e.g. look at details of associations for gene ACP1
  • In section Variation set HGDP Allele Freq to pack
    • Shows posotions were people differ from each other
    • after clicking on a particular position you get a world map with distribution of variant frequencies in different human populations
  • Browser also contains tracks displaying genomes of specific people (e.g. Jim Watson) or ancient humans (Neandertals, Denisovans)

Work with tables, downloading data

Genome browser is nice for manual browsing but also allows programmers to download data

  • each track based on one or several tables in an SQL database
  • you can download genomic sequences and data from these tables [2]
  • you can also write queries for a public SQL server [3] or create queries using Table browser forms (blue bar: Tools->Table browser)
  • conversely, you can also display your own data in "custom tracks" of the browser

Table browser examples

  • Basic type of query: e.g. export all genes in the part of the genome displayed in the browser
  • Several output formats, e.g.:
    • sequence: file of protein or DNA sequences of these genes (various settings)
    • GTF: coordinates of genes and their exons
    • Hyperlinks to genome browser: list of genes with links to the browser for each gene
    • Instead of expoert we can get summary statistics (number of items, how much sequence they cover)
  • More complex query, "intersection" of two tables: e.g. all genes that are more than 50% covered by simple repeats

Phylogenetic trees, mobyle portal

Preparing data

  • Skip this part, download the result here: http://compbio.fmph.uniba.sk/vyuka/mbi-data/cb06/cb06-aln.fa
  • UCSC browser allows us to download multiple alignments of individual genes (DNA or protein sequences)
  • In UCSC browser find gene PDE7B (phosphodiesterase 7B)
  • In the blue bar choose Tools->Table browser, track RefSeq genes, select Region: position, and Output fomat: CDS FASTA alignment and press Get output
  • At the next screen select show nucleotides. From primates select chimp, rhesus, tarsier, from other mammals mouse, rat, dog, elephant and from other species opposum, platypus, chicken, lizard, press Get output.
  • Output store on a file, remove common prefix NM_018945_ from sequence names, or completely rewrite species names

Building tree

  • We will build the tree using tools at http://mobyle.pasteur.fr/cgi-bin/portal.py
  • We will use program quicktree, neighbor joining method, bootstrap 100
    • Bootstrap means the program does 100 replicates with random subsets of the data and show how many of them contain each edge
    • Low bootstrap value means there is not enough evidence in the data for a particular branch of the tree
  • To display the tree you can use display plugins or send the tree to other display tools (button futher analysis, first choose tool in the menu)
  • "Correct tree" [4] in Conservation track settings in the UCSC browseri (based on Murphy WJ et al. Resolution of the early placental mammal radiation using Bayesian phylogenetics. Science. 2001 Dec 14;294(5550):2348-51.)
  • Our tree exhibits long branch attraction (bad position of rodents with a long branch as well as the elephant, which might be caused by sequencing errors).
  • Other programs you can try at mobyle
    • phyml: phylogenetic trees by maximum likelihood (you can choose details of the model, bootstrap, type of local moves in hill climbing,...)
    • dnapars and protpars for parsimony
    • multiple alignment by clustalw or muscle

Gene expression data

NCBI Gene Expression Omnibus http://www.ncbi.nlm.nih.gov/geo/

  • Database of gene expression data at NCBI
  • Enter GDS2925 to the search box
  • You should get Various weak organic acids effect on anaerobic yeast chemostat cultures
  • You can see basic data, such as citation, technology platform
  • Link "Expression profiles" shows plots for individual genes
  • For each gene we can get its profile neighbors - genes with similar expression
  • Data analysis tools, part Cluster heatmaps, K-means, shows results of K-means clustering for different values of K

Sequence motifs, program MEME

  • Program MEME gets a group of sequences and find a motif they have in common
  • Based of EM algorithm and probabilistic models
  • Ak mame skupinu sekvencii, mozeme hladat motiv, ktory maju
  • Go to http://meme.nbcr.net/ select MEME tool
  • As "actual sequences" enter this data
  • If the server computes too long, you can see precomputed results here

Genes and proteins