bioinformatics quiz 1

origins of bioinformatics

earliest foundations (1950-1970) focused primarily on protein sequence analysis

comprotein

first known bioinformatics software (early 1960s)

developed by Margaret Dayhoff

designed to assemble whole protein sequences (de novo) from small Edman peptide fragments

paradigm shift (1970-1980)

when bioinformatics began shifting its focus from protein analysis to DNA analysis after sanger sequencing was invented

needleman-wunsch (1970)

developed the first dynamic programming algorithm for performing pairwise protein sequence alignments

homology: orthology

homology resulting from a speciation event

defined by walter m. fitch (1970)

dayhoff/pam matrix

developed the first probabilistic model of amino acid substitutions (point accepted mutations) in 1978, using probability to measure evolutionary change

de novo sequencing

the determination of a full-genome sequence without using a known template or reference sequence

massively parallel

multiple processors working simultaneously

multiplexing

combining multiple inputs/samples into a single sequence run

overfitting

when a model built on training data shows high accuracy but significantly decreased accuracy when applied to separate validation data, indicating the model is too specific to the initial dataset features

sanger (dideoxy)

long reads (~600-1000 bp)

low throughput, typically single samples

quality loss at the beginning and end

based on chain-terminating dideoxynucleotides (ddNTPs)

illumina (MiSeq)

short reads (100-300 bp)

high/massively parallel throughput

high accuracy

bridge amplification/sequencing by synthesis where fragments attached to a flow cell are amplified into clusters

oxford nanopore (minION)

ultra long reads

high throughput, portable

moderate error rate

DNA passes through a nanopore, changes in electrical current are decoded into the DNA sequence (basecalling)

fastq file

file format that incorporates both the nucleotide sequence and associated quality scores

phred score (q)

measure of sequence quality determination

Q20 = probability of less than 1% error per base, meaning 99% accuracy

Q30 = 99.9% accuracy

coverage

the average number of reads that align to, or “cover,” known reference bases

50x genome coverage is recommendeds

single-end reads

sequence in one direction of the fragment

paired-end reads

report sequences from both directions of a DNA fragment