nj-tree-from

NJ tree from arm-level Jaccard distance matrix

Metadata

Statusdone
Assignedagent-761
Agent identity3184716484e6f0ea08bb13539daf07686ee79d440505f1fdf2de0357707034c3
Created2026-05-06T00:34:50.610349664+00:00
Started2026-05-06T00:37:06.192170070+00:00
Completed2026-05-06T00:49:45.801134119+00:00
Tagsbog-this-week, figure, nj-tree, eval-scheduled
Eval score0.93
└ blocking impact0.90
└ completeness0.95
└ coordination overhead0.95
└ correctness0.95
└ downstream usability0.90
└ efficiency0.85
└ intent fidelity0.84
└ style adherence0.95

Description

Compute neighbor-joining tree from existing 41x41 arm-level Jaccard distance matrix and produce annotated PDF. INPUTS: /moosefs/guarracino/HPRCv2/PHR_III/similarity/hprcv2.1Mb.subtelo.arm_dist_matrix.tsv (41x41); paper_prep/synthesis/ABSTRACT.md (named clades to annotate). PIPELINE: R/ape::nj() on the distance matrix; root unrooted tree at acrocentric short-arm clade if recoverable; bootstrap 1000 reps if feasible inside ~10 min agent budget else skip. ANNOTATIONS: highlight Xp/Yp via PAR1, Xq/Yq via PAR2, acrocentric short arms, 10p-18p, the 22q,21q,19q,1q,13q,17q tight clade, 4q-10q DUX4. OUTPUT FILES (under paper_prep/figures/nj_tree_arms/): nj_tree.newick, nj_tree.R (the script), nj_tree_annotated.pdf, nj_tree_annotated.png, README.md (one paragraph: which abstract clades recovered, agreement with Leiden k=15 partition, bootstrap support if computed). ACCEPTANCE: all 4 files plus the .R; PDF visibly labels each named clade; one commit msg "feat: NJ tree from arm-level Jaccard matrix with abstract clade annotations"; wg artifact records all outputs. NO SBATCH needed. NO modification to existing dist matrix.

Depends on

Required by

Log