Introduction to bioinformatics for DNA and RNA sequence analysis (IBDR01)
This course will be delivered by Malachi and Obi Griffith in Glasgow City (Scotland) Centre from 29th October - 2nd November 2018.
Analysis of high throughput genome and transcriptome data is major component of many research projects ranging from large-scale precision medicine efforts to focused investigations in model systems. This analysis involves the identification of specific genome or transcriptome features that predispose individuals to disease, predict response to therapies, influence diagnosis/prognosis, or provide mechanistic insights into disease models. During this course (PMBI01), students will perform an example end-to-end bioinformatics analysis of genome (WGS and Exome) and transcriptome (RNA-seq) data. Students will start with raw sequence data for a hypothetical case, learn to install and use the tools needed to analyze this data on the cloud, and visualize and interpret results. After completing the course, students should be in a position to (1) understand raw sequence data formats, (2) perform bioinformatics analyses on the cloud, (3) run complete analysis pipelines for alignment, variant calling, annotation, and RNA-seq (transcriptome analysis approaches will be a major component of the workshop), (4) visualize and interpret whole genome, exome and RNA-seq results, (5) leverage the identification of passenger variants for immunotherapy applications, and (6) begin to place these results in a clinical context by use of variant knowledgebases. The data, tools, and analysis will be most directly relevant to human genomics and bioinformatics research. However, many of the skills and concepts covered will be applicable to other human diseases and model organisms. Furthermore, many analysis concepts covered during the workshop will be broadly applicable to other "big data" research problems. All course materials (including copies of presentations, practical exercises, data files, and example scripts prepared by the instructing team) will be provided electronically to participants.
This workshop is primarily aimed at researchers and technical workers with a background in biology who want to learn fundamental bioinformatics skills for genomics with a particular emphasis on biomedical research applications. The course is essentially a crash course in bioinformatics for next generation sequence data analysis. It would also be useful for students with a computational background who seek an introduction to genomics technology and analysis approaches. In general, it is suitable for anyone working with genome or transcriptome (RNA-seq) data in the context of disease research. Attendees are encouraged to bring their own data or project outlines for discussion. Some time during the course will be dedicated to consultation with a team of instructors from the McDonnell Genome Institute.
Monday 29th - Classes from 09:30 to 17:30
Session 1. Introduction to genomics and bioinformatics.
In this session, students will be introduced to key concepts of genomics and their application to genomics research and precision medicine in cancer. An introduction to next-generation sequencing platforms and related bioinformatics approaches will also be provided.
Core concepts and tools introduced: fundamentals of genome and transcriptome analysis, next-generation sequencing, precision/personalized medicine approaches (using cancer as an exemplar disease).
Session 2. Introduction to genomics data, file formats, QC, and cloud analysis.
In this session, students will be introduced to a hypothetical patient case and related samples to be analyzed throughout the course. Students will be provided with an introduction to the whole genome, exome, transcriptome and other data sets we have generated for this test case. Information on where to get the raw data and how to access it (and other test data) will be provided. Using this data as an example, the students will learn fundamentals of next generation sequence (NGS) data formats. The students will also be introduced to accessory files needed for analysis including reference genomes, reference transcriptomes, and annotation files. Tools for QC analysis of raw data will be demonstrated. Since most analysis will be performed on the cloud, each student will learn how to launch and log into their own cloud compute environment. Students will learn how to install bioinformatics tools and learn to use some of the most broadly useful tool kits for NGS data.
Core concepts and tools introduced: file formats (Fasta, FastQ, SAM/BAM/CRAM, VCF, GTF), bedtools, Picard, samtools, fastQC, cloud computing (AWS, EC2).
Tuesday 30th - Classes from 09:30 to 17:30
Session 3. Primary genome data analysis (sequence alignment and visualization).
In this session, we will start to complete analysis of NGS data at the command line. Students will log into the cloud, and starting with their own copy of the raw data will align the whole genome and exome data to a reference genome. Following alignment, students will conduct a second quality analysis of the data and learn to visualize alignments in IGV.
Core concepts and tools introduced: alignment algorithms, reference indexes, BWA, BWA-mem, alignment indexes, alignment flags, genome browsers, duplicate marking, alignment merging and sorting, IGV.
Session 4. Whole genome and exome variant calling and annotation.
In this session, we will introduce different algorithms for identifying sequence variations of various types from either whole genome or exome data (or both). Both germline and somatic variant calling will be covered. For each, students will learn strategies for identifying false positives and increasing confidence in individual predictions by manual or secondary examination of the alignments. Variant types detected will include single nucleotide variants (SNVs), small insertions and deletions (indels), copy number variants (CNVs) and structural variants (SVs). Students will learn strategies for visualizing and presenting variants of each type. After producing filtered variant results of each type, annotation methods and resources relevant to each variant type will be demonstrated.
Core concepts and tools introduced: germline variation, somatic variation, variant calling, false positives, false negatives, alignment artifacts, manual review, svviz, manta, GATK, Strelka, MuTect, VarScan, CopyCat, Lumpy.
Wednesday 31st - Classes from 09:30 to 17:30
Session 5. RNA-seq analysis (introduction, alignment and abundance estimation).
In this session, students will learn about fundamentals of RNA-seq data analysis and perform initial QC and alignment of the raw transcriptome data. Appropriate sample comparisons for RNA-seq and other experimental design and analysis considerations will be discussed in detail.
Core concepts and tools introduced: reference transcriptomes, normalization, batch effects, replicates, spliced alignment algorithms, RNA-seq data trimming, RNA assembly algorithms, RNASeqQC, HISAT, StringTie.
Session 6. RNA-seq analysis (fusions, differential expression, and
The uses of transcriptome data in biological research are remarkably varied. Students will pursue several strategies in this section. Fusion detection, an RNA-seq specific variant detection approach will be performed. The expression abundance results from the previous section will be used to identify a list of highly expressed genes. Comparison to RNA-seq data from a cohort of related samples will be used to identify expression outliers. Expression clustering algorithms will be used to stratify our case using a known expression signature. More advanced classification and pathway based approaches to stratification will be briefly introduced.
Core concepts and tools introduced: fusion calling, outlier analysis, expression clustering, stratification, heatmaps, Ballgown, pizzly.
Thursday 1st - Classes from 09:30 to 17:30
Session 7. Prioritization, visualization and interpretation.
In this session, students will learn about procedures for refining the final results obtained from the previous analyses of our case data. Genome and transcriptome variant observations will be prioritized according to various annotation strategies. These vary from algorithmic predictions of pathogenicity to intersecting with results from population databases. Students will also learn how to integrate results from the DNA and RNA-seq analyses. For example, variants will be prioritized according to their expression status, allele specific expression bias, and the abundance of associated genes. Fusions predicted in the RNA will be confirmed in the DNA. Visualization techniques will be used to place variant observations from our case in the context of a cohort of previously sequenced cases with the same disease. A group discussion will tackle how to approach creating a final clinical interpretation for our example patient.
Core concepts and tools introduced: allele specific expression, clonality, GenVisR, gnomad, CADD, bam-readcount, integrate.
Session 8. Gene/variant knowledgebases and clinical actionability.
In this session, students will learn the fundamentals of interpreting genome and transcriptome observations in a clinical context. The final candidate observations for our example case will be examined using various clinical interpretation tools and databases.
Core concepts and tools introduced: Druggability, actionability, sensitivity, resistance, predictive variants, diagnostic variants, prognostic variants, predisposing variants, the ACMG and AMP guidelines for clinical actionability, variant knowledgebases, CBioPortal, CIViC, ClinVar, DGIdb, PharmGKB.
Friday 2nd - Classes from 09:30 to 16:00
Session 9. Leveraging passenger variants (monitoring and immunogenomics).
Up until this point, we have been focused on identifying, annotating and interpreting variants that are potentially relevant to disease biology or clinical interpretation. These are variants that are deemed functional, actionable, or of some known clinical relevance. What about those variants that may be unusual or unique to this case but of no known significance? What about the "passenger" variants? In this section, we will explore two broad strategies that leverage passenger variants in a clinically useful way (using cancer as an exemplar disease for this approach). First, we will examine their potential use in tracking response to therapy. Second, we will explore the possible immunogenomic implication of passenger variants by designing a personalized cancer vaccine for our example case.
Core concepts and tools introduced: cfDNA, serial analysis, immunotherapy, pVacTools.
Session 10. Application to your own data
Optional free afternoon to cover previous modules or consult with the team of instructors. In this session, students will be free to work on their own, or in groups on the previously covered sections. Furthermore, students can consult with the team of instructors on their own experiments or get practical advice for analyzing their own data. Our hope is to make this session as interactive and useful as possible.
To learn more about the team of instructors, please visit www.griffithlab.org and http://genome.wustl.edu/people/groups/detail/griffith-lab/.
Email firstname.lastname@example.org with any questions
Check out our sister sites,
- www.PRstatistics.com (Ecology and Life Sciences)
- www.PRinformatics.com (Bioinformatics and data science)
- www.PSstatsistics.com (Behaviour and cognition)
Oliver Hooker PhD.
2018 publications -
Phenotypic and resource use partitioning amongst sympatric lacustrine brown trout, Salmo trutta. Biological Journal of the Linnean Society. DOI 10.1093/biolinnean/bly032
6 Hope Park Crescent
Edinburgh EH8 9NA