MegaMiner: Searching TESS Threshold Crossing Events for Anomalous Signals

Sectors processed: 36–80. Last updated: 2026-09-16

MegaMiner pipeline1TESS TCEsSPOC transit detections2Anomaly rankingleast planet-like first3Remove known EBsvetted EB catalogs4Diagnostic flagsSPOC DV tests5ML classifiersin development6Anomaly candidatesfor expert review
260,540
total TCEs
131,743
unique signals
107,508
unique TIC IDs
11,798
unflagged by diagnostics
116,099
unflagged by vetted catalogs
10,041
total unflagged survivors

Gaia Stellar Parameters

Host-star parameters from the Doyle et al. (2024) TESS–Gaia cross-match.

Gaia cross-match (valid stellar parameters)
71,119 / 131,743 (54.0%)
Tmag in range (6–14)
126,158 / 131,743 (95.8%)
Surface gravity (log g ≥ 3.5)
129,122 / 131,743 (98.0%)
Parallax S/N (≥ 5)
70,947 / 131,743 (53.9%)
RUWE (< 1.4)
57,445 / 131,743 (43.6%)

Diagnostic flags

Flags derived from the diagnostic tests in the SPOC Data Validation (DV) reports for each TCE, with thresholds set by MegaMiner. A signal with any of these flags is not counted as an unflagged survivor.

Suspected eclipsing binary (SPOC)
7,268 / 131,743 (5.5%)
Transit fit did not converge
13,246 / 131,743 (10.1%)
Invalid odd/even statistic
22,624 / 131,743 (17.2%)
Background / blended EB (ghost ratio < 1)
72,136 / 131,743 (54.8%)
Off-target centroid offset (> 3σ)
57,551 / 131,743 (43.7%)
Matching-period signals (within 0.01 d)
46,483 / 131,743 (35.3%)
Large odd/even depth difference (> 35)
5,408 / 131,743 (4.1%)
Low S/N (< 10)
53,340 / 131,743 (40.5%)
Significant secondary eclipse (weak-secondary ≥ 7)
21,201 / 131,743 (16.1%)

Catalog flags

Vetted eclipsing-binary catalogs (True = known EB). Vetted membership gates; unvetted is annotation only.

Prša et al. (2022) Vetted EB Catalog
5,634 / 131,743 (4.3%)
Kostov et al. (2025) Vetted EB Catalog
1,152 / 131,743 (0.9%)
Oddo et al. (2025) M+M EB Catalog
134 / 131,743 (0.1%)
Shan et al. (2025) CALNet EB Catalog (ML-derived, 2-min)
9,119 / 131,743 (6.9%)
Combined Catalog EBs
15,644 / 131,743 (11.9%)
Kostov et al. (2025) Unvetted Candidates (annotation only)
14,636 / 116,099 (12.6%)

Distributions

Transit depth (ppm)
014,16128,3222.8e-071.6e-039.35.4e+043.1e+08value (log scale)count
n=116,099 · median 3,927 · range 0 to 3.14e+08
Orbital period (days)
09,79619,5910.20.984.924.2120value (log scale)count
n=116,099 · median 7.601 · range 0.2 to 119.9
Reduced chi-squared
053,998107,9970.049.2e+042.1e+114.7e+171.1e+24value (log scale)count
n=116,099 · median 0.9451 · range 0.04286 to 1.074e+24
Model fit SNR
037,93275,8641.0e-073.2e-0399.33.1e+069.6e+10value (log scale)count
n=116,099 · median 12.59 · range -1 to 9.59e+10
TESS magnitude
09,55619,1131.66.411.216.020.9valuecount
n=116,099 · median 10.85 · range 1.609 to 20.86
Effective temp (K)
030,70561,4102,7211.6e+042.9e+044.2e+045.5e+04valuecount
n=116,099 · median 5,780 · range 2,774 to 5.542e+04

Flag co-occurrence

Suspected EB (SPOC)
Fit not converged
Invalid odd/even
Background EB (ghost)
Centroid offset
Matching period
Large odd/even
Low S/N
Secondary eclipse
Catalog EB
Suspected EB (SPOC)7,2687,2687,2683,6344154,6591077,2682,6262,227
Fit not converged7,26813,24613,1936,9381,9806,51627013,1623,7132,960
Invalid odd/even7,26813,19322,62411,7694,4879,37547116,4485,1754,495
Background EB (ghost)3,6346,93811,76972,13641,50624,5581,69132,48510,5006,024
Centroid offset4151,9804,48741,50657,55118,15267723,5496,5962,879
Matching period4,6596,5169,37524,55818,15246,4832,22917,20513,93310,638
Large odd/even1072704711,6916772,2295,4089434801,381
Low S/N7,26813,16216,44832,48523,54917,20594353,3407,4705,212
Secondary eclipse2,6263,7135,17510,5006,59613,9334807,47021,2013,838
Catalog EB2,2272,9604,4956,0242,87910,6381,3815,2123,83815,644

Cell = number of signals tripping both flags. Row shading is relative to the row flag's own total (diagonal).

Preliminary results

The 10,041 unflagged survivors from sectors 36–80 are the signals that pass every diagnostic and catalog check above. They are ranked by anomaly score and form our candidate list for follow-up vetting. The list is part of our preliminary, unpublished results and is available to collaborators and sponsors on request.

Request access to preliminary results

Access is granted individually by the project team. Once approved, you can view the full ranked candidate list, which is updated as new TESS sectors are processed. Please use the same email address on the form that you will use to log in.