Skip to content

Latest commit

 

History

History

galaxy_rna-seq_tuxedo

README.md

Title: The Cloud and the Shell - Applied Bioinformatics on the Example of Gene Expression Analysis using Unix and freely available Open Source Tools
Author: Alexander Kerner
EMail: training at silico-sciences.com
Seminar Ruprecht-Karls-Universität Heidelberg

Applied Bioinformatics in the Cloud and in the Shell

RNA-Seq Analysis Using Unix and Open Source Tools

[TOC]

RNA-Seq using Galaxy, IGV and the Tuxedo Suite

Getting the Data

For this tutorial we will work with the following dataset:

http://www.ebi.ac.uk/ena/data/view/ERP003613

http://www.ebi.ac.uk/ena/data/view/ERS326931-ERS327025

NGS data processing is usually very time consuming, which is why some data has been already partically processed.

Furthermore, the data has been reduced in size (chromosome 3 only) so processing will not take too long (usually a matter of hours or even days).

Below you find links to a subset of the original data, the preprocessed data as well as the chromosome 3 only datasets.

Lung Original

ID URL md5 Read Count
ERR315424 http://www.ebi.ac.uk/ena/data/view/ERS326948 74368f3aaf69fa32a73c43e6cf7c32ee 083f9c11f8c65a40bff289943d49524f 8,761,137
ERR315353 http://www.ebi.ac.uk/ena/data/view/ERS326952 e07815132f75ad4acd626ba017b0f1b3 ad9a286773704d120b84558253f915b1 13,576,003
ERR315487 http://www.ebi.ac.uk/ena/data/view/ERS326952 64f66f5dfe6a15b72d060fe30900f899 c07809ef2d7e905dddf718d6df17ebea 13,669,047
ERR315444 http://www.ebi.ac.uk/ena/data/view/ERS327010 a47bb31a6beb7a59a9085313c824627e c704f62abbb76c7bb1408892f7f6716a 6,387,998

Lung Reduced

ID URL md5 Read Count
ERR315424 ftp://public@silico-sciences.com/new/ERR315424/ERR315424_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315424/ERR315424_chr03_2.fastq.gz 2d8e2cfce8449851eb4a9c0a39c417cc 3348c4295275b31136e800bdefe8ebee 3873
ERR315353 ftp://public@silico-sciences.com/new/ERR315353/ERR315353_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315353/ERR315353_chr03_2.fastq.gz 32d6680538b3f7adf76e33810cef3864 1b5c65d9b826a24b328f1fcb4551edb1 3200
ERR315487 ftp://public@silico-sciences.com/new/ERR315487/ERR315487_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315487/ERR315487_chr03_2.fastq.gz 2395a015534e03cd1e842275804e26f8 a67e50045f35d013f75223449978fdfe 4875
ERR315444 ftp://public@silico-sciences.com/new/ERR315444/ERR315444_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315444/ERR315444_chr03_2.fastq.gz c75154f04224905db8716641127b3363 bb731b43a4c922dd3b92bd711a35064d 3276

Stomach Original

ID URL md5 Read Count
ERR315369 http://www.ebi.ac.uk/ena/data/view/ERS326950 60b67df94f041c510977575525463a9d fe6230895aa051839e5ada69a8a0e5c9 16,004,990
ERR315485 http://www.ebi.ac.uk/ena/data/view/ERS326950 6d8f45845c9dbb61e8d7a40db99e2c9b f1117143c7b450c576ade3dd32ba6df0 15,684,423
ERR315467 http://www.ebi.ac.uk/ena/data/view/ERS327000 ac54c6fef08daf49b1c339f88231494c 233f5b072fc327ff1caa97045feddc50 20,097,530
ERR315379 http://www.ebi.ac.uk/ena/data/view/ERS327014 083bc3884d96c57cab5d2561a37e6304 c97fbd4b8459382ae35cea6b86ffbe03 22,610,772

Stomach Reduced

ID URL md5 Read Count
ERR315369 ftp://public@silico-sciences.com/new/ERR315369/ERR315369_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315369/ERR315369_chr03_2.fastq.gz f4c800960c694f32605ba2a2b64a0a35 19d06f57d1577be2900cf54af1e95b88 15182
ERR315485 ftp://public@silico-sciences.com/new/ERR315485/ERR315485_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315485/ERR315485_chr03_2.fastq.gz 2abf1404b10fd5cbd942212ef7daa2a7 43b4693d8934d587c345d48d38f7eee0 14775
ERR315467 ftp://public@silico-sciences.com/new/ERR315467/ERR315467_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315467/ERR315467_chr03_2.fastq.gz 6674b87cc3397d4e02dfe4a5423c6cff 0c16723ca8afc2b43d59ea6d2e239216 14893
ERR315379 ftp://public@silico-sciences.com/new/ERR315379/ERR315379_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315379/ERR315379_chr03_2.fastq.gz 553d656b886e51d0aeda26d82378fd5c d5797e2e2f9673cb5c8af6dadd7aa93f 14812

Heart Original

ID URL md5 Read Count
ERR315384 http://www.ebi.ac.uk/ena/data/view/ERS326973 0c13852e9bbddc1c2b056745ee6ffe6c 0dcdbe173ed117fff7995ffaa0b4f1b6 16,695,109
ERR315413 http://www.ebi.ac.uk/ena/data/view/ERS326973 f8473591adbcb929ad729d1427b97d52 373d9a72ba830a96504c62a1486d55c1 16,596,950
ERR315356 http://www.ebi.ac.uk/ena/data/view/ERS327004 db1a5470bd92049dfc7af541e17ad0ad 090b337910da0209f819650e7565da02 17,732,535
ERR315430 http://www.ebi.ac.uk/ena/data/view/ERS327004 05e09085cc4e09b3626e35c68e8284fe 8b80a1c1ec49e42b08d0dad195dd9113 17,381,370

Heart Reduced

ID URL md5 Read Count
ERR315384 ftp://public@silico-sciences.com/new/ERR315384/ERR315384_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315384/ERR315384_chr03_2.fastq.gz 3ddef3455381ecd4099493055253c455 5bfe358c01bdfe43ec8025e60e85fd53 6683
ERR315413 ftp://public@silico-sciences.com/new/ERR315413/ERR315413_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315413/ERR315413_chr03_2.fastq.gz 5a11a2924f7d452722756f4e42ddc9ef 298941f7572f423302d5ae13bf7cd0eb 6766
ERR315356 ftp://public@silico-sciences.com/new/ERR315356/ERR315356_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315356/ERR315356_chr03_2.fastq.gz 8a8dffabeca2836ace01f0a4e8de9434 598a7f2ead01aa65a2a4372837cd9d2f 8732
ERR315430 ftp://public@silico-sciences.com/new/ERR315430/ERR315430_chr03_1.fastq.gz ftp://public@silico-sciences.com/new/ERR315430/ERR315430_chr03_2.fastq.gz 2c6e035f6c1b7028e5b1a5eeba1571b2 47194c4f4a349e9f5f65b6734000d818 8496

Find the main galaxy server here

https://usegalaxy.org/.

If you experience any problems you can try as well another server, as there are many available.

Import Data into Galady

upload

Galaxy provides different possibilities to import/ upload data.

Here, you can copy&paste the file URLs into the Galaxy upload wizard to transfer the data from server to server, without the need to download it to your local machine.

Reference Sequence

Use the link below to get the reference sequence:

ftp://public:public@silico-sciences.com/new/3.fa.gz

Reference sequences are stored in the FASTA format.

The file extension for FASTA files is .fasta, .fa or the according compressed version fasta.gz and fa.gz.

Reference Annotation

Use the link below to get the reference annotation:

ftp://public:public@silico-sciences.com/new/genes_chr03.gtf.gz

The file format should be GTF.

This file provides meta information on the (anonymous) reference sequence, such as exons, CDSs or start- and stop codons.

Sequencing Reads

Get the "chr03-only" fastq.gz files.

NGS data is stored in the FASTQ format. These files are usually the starting point of the NGS data processing.

If the reads are paired-end, you are usually provided with two files per sample (_1.fastq.gz and _2.fastq.gz).

Inspecting the Data and do Some Quality Control

  1. Take a look at the content of the uploaded files.

    view

  2. Use FastQC to take a look at the overall read quality and sample details. The quality scores in the FASTQ files can be in different encodings. FastQC can automatically detect the encoding and reports it in the "Base Statistics" sections:

    base-statistics

    The tools we are using need to know about the quality encoding that is present in the FASTQ files to probably interpret the quality of the base calls. Our files are in "Sanger" format, so we need to edit the data attributes for the loaded fastq.gz files to "fastqsanger".

    Find out more on handling FastQ quality scores in Galaxy here.

  3. Use Trim Galore! to trimm the reads.

  4. Compare read quality of trimmed vs. non-trimmed reads.

  5. Answer the following questions for trimmed and non-trimmed data:

    1. Which quality encoding was used?

    2. What is the read length distribution?

    3. What is the quality score distribution?

Mapping Reads to the Reference Sequence

  1. Use TopHat to map reads to reference from history (3.fa.gz). TopHat is a splice-aware aligner, so it can handle RNA-Seq data and is able to align reads across introns.

Bowtie is not suitable for all sequence alignment tasks. It does not allow alignments between a read and the genome to contain large gaps; hence, it cannot align reads that span introns. TopHat was created to address this limitation. TopHat uses Bowtie as an alignment 'engine' and breaks up reads that Bowtie cannot align on its own into smaller pieces called segments. Often, these pieces, when processed independently, will align to the genome. When several of a read's segments align to the genome far apart (e.g., between 100 bp and several hundred kilobases) from one another, TopHat infers that the read spans a splice junction and estimates where that junction's splice sites are. (Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks)

  1. Choose Paired-end (as individual datasets).

  2. Select forward and reverse reads from each sample.

  3. Choose Use a genome from history.

  4. Select 3.fa.

  5. Take a look at the newly added files (BAM).

  6. Answer the following questions:

    1. How many reads could be sucesfully aligned?

    2. How many reads could be uniquely aligned?

Other read mapper:

  1. HISAT2
  2. STAR
  3. GSNAP
  4. MapSplice

Use IGV to Visualize the Mapping

  1. Download or web-start IGV.

    Ubuntuusers: you can find IGV in the repositories. To install, type sudo apt-get install igv.

  2. Download mappings accepted_hits.bam and its index file.

  3. Load bam files into IGV.

    1. Navigate by typing gene names into the search box.

    2. Navigate by Drag&Drop.

    3. Zoom in and out using the "railroad track".

    4. Double click to get online info.

    5. Change display:

      1. Collapsed

      2. Expanded

      3. Squished

    6. Load additional data from server, e.g. dbSNP.

    7. Find out what else you can do!

Other Genome Browsers:

  1. UCSC
  2. Ensembl

Calculate Gene Expressions

Use the reduced datasets to get an idea of the workflow and the input and output files.

Use the original dataset to perform a real gene expression comparison between lung, stomach and heart.

  1. Use Cuffquant to precompute gene expression levels.

    From the Cufflinks manual:

    Quantifying gene and transcript expression in RNA-Seq samples can be computationally expensive. Cuffquant allows you to compute the gene and transcript expression profiles and save these profiles to files that you can analyze later with Cuffdiff or Cuffnorm. This can help you distribute your computational load over a cluster and is recommended for analyses involving more than a handful of libraries.

  2. Use Cuffnorm to create normalized expression values.

    From the Cufflinks manual:

    Sometimes, all you want to do is normalize the expression levels from a set of RNA-Seq libraries so that they're all on the same scale, facilitating downstream analyses such as clustering. Expression levels reported by Cufflinks in FPKM units are usually comparable between samples, but in certain situations, applying an extra level of normalization can remove sources of bias in the data. Cuffnorm normalizes a set of samples to be on as similar scales as possible, which can improve the results you obtain with other downstream tools.

  3. Use Cuffdiff to find significant changes in expression level.

    From the Cufflinks manual:

    Comparing expression levels of genes and transcripts in RNA-Seq experiments is a hard problem. Cuffdiff is a highly accurate tool for performing these comparisons, and can tell you not only which genes are up- or down-regulated between two or more conditions, but also which genes are differentially spliced or are undergoing other types of isoform-level regulation.

    1. Omit Tabular Datasets: No.
    2. Generate SQLite: Yes.

  4. Use cummeRbund to visualize data.

Data Analysis

  1. Download gene_differential_expression_testing.tabular (gene_exp.diff).

  2. Open the file with your favourite spreadsheet software.

    1. Pay attention to correct importing of numerical values (correct decimal delimiter).

    2. Filter rows with non-numerical values for fold changes (-inf and inf).

    3. Sort table by p-value and absolute value of log2(fold_change).

      Hint: Insert an extra column that contains absolute exp. values (use the abs function).

  3. Answer the following questions:

    1. Which gene shows the strongest fold change (log scale and linear scale)?

    2. Which gene shows the strongest significant fold change (log scale and linear scale)?

    3. Which gene shows the weakest fold change (log scale and linear scale)?

    4. Which gene shows the weakest significant fold change (log scale and linear scale)?

Back to Index

References

  1. Galaxy Wiki

  2. Galaxy forum

  3. FASTQ format#Quality

  4. SAM/BAM format

  5. Paper Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks

  6. Paper Analysis of the Human Tissue-specific Expression by Genome-wide Integration of Transcriptomics and Antibody-based Proteomics

  7. Cufflinks manual

  8. CummeRbund manual

  9. FPKM explained