Statistical Significance & Scoring Matrices — Hands-on

Gain expertise in statistical significance and sequence scoring matrices through real-world, benchmark-driven practical sessions.

Statistical Significance & Scoring Matrices

Session Index

Session 1

Fee: Rs 6300

Significance: P, E & Null Models

P-value vs E-value; extreme value distributions (Gumbel)

Theory

Null models, shuffling, background frequencies & λ/K

Score distSimulations

Interpreting significance across databases & search sizes

ScalingCutoffs

Session 2

Fee: Rs 8400

BLOSUM/PAM & Gap Penalties

PAM series vs BLOSUM family: construction & calibration

PAMBLOSUM

Affine/convex gap models & score normalization

Gap open/extendZ-scores

Choosing parameters for sensitivity vs specificity

Grid searchHeuristics

Session 3

Fee: Rs 11200

FDR, Resampling & Curves

Multiple testing: Bonferroni, BH/FDR, q-values & π0

FDRq-value

Permutation & bootstrap for robust significance

ResamplingConfidence bands

ROC/PR curves, calibration plots & threshold tuning

ROC/AUCPR-AUC

Session 4

Fee: Rs 14000

Mini Capstone: Benchmarks & Reports

Design benchmark on curated datasets & gold standards

Theory

Compare BLAST/DIAMOND/MMseqs2 settings & matrices

BLAST+ DIAMOND MMseqs2

Publish-ready figures & narrative with stats appendix

pandas matplotlib JupyterLab

PDF

Statistical Significance & Scoring Matrices Training | E-values, BLOSUM/PAM, FDR & Benchmarks