HiPiler: Visual Exploration Of Large Genome Interaction Matrices With Interactive Small Multiples

Slide 1

Slide 1 text

HiPiler Visual Exploration of Large Genome Interaction Matrices with Interactive Small Multiples Fritz Lekschas, Benjamin Bach, Peter Kerpedjiev, Nils Gehlenborg, and Hanspeter Pﬁster

Slide 2

Slide 2 text

3 million × 3 million

Slide 3

Slide 3 text

> 10.000 pattern instances but small total size

Slide 4

Slide 4 text

> 10.000 pattern instances but small total size How can we explore and compare many local patterns in this very large matrix?

Slide 5

Slide 5 text

HiPiler

Slide 6

Slide 6 text

Social Networks Fans, connectors, and cliques Computer Networks Bottlenecks and hubs Gene Networks Feed-forward loops Giga-pixel Images Pattern recognition

Slide 7

Slide 7 text

Structure of the Genome Acknowledgements: N. Abdennur, B. Alver, H. Belaghzal, A. van den Berg, J. Dekker, G. Fudenberg, J. Gibcus, A. Goloborodko, D. Gorkin, M. Imakaev, Y. Liu, L. Mirny, J. Nübler, P. Park, H. Strobelt, and S. Wang.

Slide 8

Slide 8 text

DNA Cell Nucleus Contact Sequencing

Slide 9

Slide 9 text

Cell Nucleus Contact Sequencing Matrix Fixed Ordering Altered DNA ordering is associated with severe diseases!

Slide 10

Slide 10 text

Challenges • Detected by algorithms • Occur frequently • "Noisy" results Goals • Quality assessment • Pattern stratiﬁcation • Pattern correlation Points Blocks

Slide 11

Slide 11 text

• How do speciﬁc pattern or average pattern look? • How variant and noisy are detected pattern? • Are there subgroups among the pattern? • How are patterns related to other data attributes? • What does the patterns neighborhood look like?

Slide 12

Slide 12 text

TECHNIQUES? • Pan & Zoom  Kerpedjiev et al.: HiGlass • Lenses / Multifocus  Rao and Card: Table Lense   Elmquist et al.: Melange • Abstraction / Aggregation  Dunne et al.: Motif Simpliﬁcation   Elmquist et al.: ZAME • Small Multiples  Bach et al.: Multipiles

Slide 13

Slide 13 text

Cut the Matrix into Pieces!

Slide 14

Slide 14 text

Cut the Matrix into Pieces!

Slide 15

Slide 15 text

Cut the Matrix into Pieces!

Slide 16

Slide 16 text

Cut the Matrix into Pieces!

Slide 17

Slide 17 text

Cut the Matrix into Pieces!

Slide 18

Slide 18 text

Cut the Matrix into Pieces!

Slide 19

Slide 19 text

HiPiler

Slide 20

Slide 20 text

HiPiler

Slide 21

Slide 21 text

1. FILTERING Assess quality & separate signal from noise

Slide 22

Slide 22 text

1. FILTERING

Slide 23

Slide 23 text

1. FILTERING

Slide 24

Slide 24 text

1. FILTERING

Slide 25

Slide 25 text

1. FILTERING

Slide 26

Slide 26 text

1. FILTERING

Slide 27

Slide 27 text

2. AGGREGATE Stratify patterns and assess pattern variability

Slide 28

Slide 28 text

2. AGGREGATE

Slide 29

Slide 29 text

2. AGGREGATE

Slide 30

Slide 30 text

3. CONTEXT Correlate patterns with each another & other pattern types

Slide 31

Slide 31 text

3. CONTEXT

Slide 32

Slide 32 text

Pile Inspection Attribute correlations Multidimensional Clustering Dataset Comparison More at http:/ /vcg.seas.harvard.edu/pubs/hipiler

Slide 33

Slide 33 text

User study with 5 domain experts: Evaluating usability and usefulness Snippet approach is useful: Average / variance assessment and parameter estimation Context matters: Coordination between the snippets and matrix is highly appreciated HiPiler is easy-to-use and useful: Domain experts ask for local installations Limitations: Fixed matrix ordering and ﬁxed aspect ratio of snippets EVALUATION

Slide 34

Slide 34 text

CONCLUSION Coordinate Aggregate Arrange & Filter Separate Explore

Slide 35

Slide 35 text

NEXT?

Slide 36

Slide 36 text

HiPiler PAPER   vcg.seas.harvard.edu/pubs/hipiler LIVE   hipiler.higlass.io CODE   github.com/ﬂekschas/hipiler