Human human reference species mouse, chimp, arabidopsis…

100 million

• 100 million

human

• human

• reference species mouse, chimp, arabidopsis…

• agricultural species cattle, sheep, pig, … rice, wheat, grape …

• bacterial disease, human “ecosystem”

Individuals

• Individuals

• Populations disease and “quantitative traits”

• Somatic and tumor genomes

• Transcriptome of child and parents

• Bacterial populations of individuals

Indexing

• Indexing

• Alignment

• SNP/MNP/Indel/SV calling

Human genome (3 billion nt)

• Human genome (3 billion nt)

• 1 billion reads of 100 nt coverage of 30

• Indexing + Aligning in 27 minutes

Indexing

• Indexing

• Alignment

Varying levels of extraction of reads across genome (use differences)

• Varying levels of extraction of reads across genome (use differences)

• Locate boundaries (as accurately as possible)

• Extract number of variants

• Use SNPs

Mapping reads back onto a database of known bacteria/viruses

• Mapping reads back onto a database of known bacteria/viruses

• Many are ambiguous

• Many don’t map at all

• Estimate frequency of each species

• Remove human “contamination”

Map reads to database

• Map reads to database

• Estimate most likely frequencies a hill climbing estimation problem

• Can anything be done about unmapped reads?