Please enable JavaScript.

Coggle requires JavaScript to display documents.

Variants, Variants - Coggle Diagram

- - - - Biological Effects
        
        If in coding region
        
        Frameshift mutations
        
        If indel is of 1 or 2 bases - affects a large stretch of protein coding sequence, can also lead to premature STOP codon.
        
        if indel is of 3 bases - affects one or two codons in any of the three reading frames of an RNA
        
        If in non-coding region
      - Examples
        
        1 bp insertion in CYP3A5*7 gene - pre-disposition to hypertension
        
        1 bp deletion in HERC2 gene - causes blue eye colour
    - - Single nucleotide substitutions
        
        If SNV found commonly in a population (germline) = SNP
      - Biological Effects
        
        If in coding region, can affect function and structure of protein
        
        Synonymous mutation
        
        Silent mutation - Change in base leads to no change in AA
        
        Non synonymous mutation
        
        Missense mutation - Base change leads to change in AA
        
        Nonsense mutation - Base change leads to stop codon
        
        Can lead to protein structure and function change
        
        Can prevent translation of prematurely terminated RNA by causing its decay - nonsense mediated decay
        
        If in non-coding region, can affect:
        
        Gene regulation, if present in UTRs
        
        Splicing, if present in intron edges (splice sites)
        
        Promoter binding, if present upstream of gene
  - - - Duplication - of a gene
        
        Can be detected using read depth difference during sequencing (NGS)
        
        Detection using NGS
        
        Breakpoint - point at which relative to the reference genome, the aligned sequence is broken and reassembled
        
        For all Structural variants
        
        Paired end mapping - when the read pairs don't align at an expected distance from each other.
        
        Gives approximate breakpoint location (somewhere between the reads of a pair), not single-base resolution
        
        More confidence because read depth can help
        
        Diagrams
        
        Lecture 11, slide 32-40
        
        Split-read mapping -- split reads in half and map them separately to reference. When the different halfs or ends map to different regions of the reference.
        
        Single-base resolution of breakpoint, but need longer reads (>200 bp) and maybe harder to get enough read depth.
        
        Tools
        
        Manta, GRIDSS
        
        De Novo assembly using reads from relevant regions
      - Deletion - of a gene
        
        Biological Effects
        
        Duplication
        
        Over expression of RNA and proteins
        
        Affects interaction between and regulation of genes
        
        Deletion
        
        Loss of heterozygosity - only one allele of a gene lost
        
        implicated in Bipolar disorder, schizophrenia, autism, etc.
        
        Can be detected using B-allele frequencies - E.g. in case of gene insertion, diploid becomes triploid, reflecting in the read (from NGS) frequency distribution
        
        Lecture 11, slide 29
      - Translocation - deletion of gene from position A and insertion of it in position B
      - Inversion - gene changing between strands
      - Chromothripsis
        
        Multiple simple events at once, when taken to extreme -- shattering and repair of chromosome or group of chromosomes
        
        Can lead to cancer
        
        Cancer - lecture 11, slide 43-67
      - Database for structural variants - Database of Genomic Variants (DGVa)
    - - Somatic CNVs - Copy Number Alterations (CNA)
- - - - Finding the probability of each genotype (e.g. AA,AB,BB), given the observed data = P(G | D)
        
        Lecture 10, 35-60
        
        Bayes theorem
        
        P(D)
        
        Constant - probability of getting the data we got, out of all possible data randomly.
        
        Important when variants can be rarer than false positives/errors
        
        P(G) - prior probability of each genotype based on knowledge of sample's biology.
        
        Each variant is not as likely as a non-variant.
        
        Variants are quite rare that we need a lot of read evidence to be confident about the variant call.
        
        P(D | G)
        
        Different binomial distributions depending on what the genotype being considered is (AA, AB or BB)
        
        Lecture 10, slide 52-55
        
        Will be more definite if there is low sequencing error and higher read depth
        
        Includes consideration of error model for sequencing tech.
    - - Number of reads representing reference allele
      - Number of reads representing alternate allele
      - base qualities
      - Strand bias
    - - Snippy - a straightforward way to run multiple alignment and variant calling (SNPs) on the command line (SNVs, indels and rearrangements)
      - FreeBayes - when given an input of BAM files and a reference, it produces an output describing
        SNPs, indels and complex variants in the alignments
      - Ensemble DataSlicer
- - - - Lecture 11, slide 22
    - - Haplotypes
        
        Haploid genotype - set of phased genotypes. We know which of the variants occur on the same chromosome.
        
        Some SNPs are inherited together due to linkage disequilibrium - SNPs are co-inherited either due to being physically close together or in recombination hotspots
        
        Use
        
        Genotype imputation - use haplotype information (international hapmap project) to deduce likely genotypes of other unmeasured SNPs, that are linked to the genotypes of measured SNPs --- not useful if the whole genome is being sequenced