RSS GitHub
⚡ Latest Lab Discoveries Recent Reports

See more Lab Discoveries below ↓

Tier:
Domain:
Topic:
Showing 10 of 50
← Return to CRV Science Labs / Molecular Systematics & Pipeline Validation
#83 🚩 Critical Report Tier 2: Validation 🧪🧪🧪🧪 4/5 • Strong Emergent Signal Molecular Systematics & Pipeline Validation

Tier 2 Dual-Pass Blinded and Dayhoff-6 Validation of Heimdallarchaeia Phylogenomics: Resolving Compositional Long-Branch Attraction and Metagenomic Bin Chimeras

Validating 24 Development and Untouched Reserve cohorts (768 kb) via de-identified SHA-256 sequence extrapolation and 6-state Dayhoff recoding to eliminate confirmation bias and catch metagenomic assembly artifacts

By Bryan P. White • 15 min read • Medium (Undergraduate Biology & Biochemistry)
✨ Executive Summary

To verify that eukaryotic-like cellular, regulatory, and ribosomal signals in Candidatus Heimdallarchaeia are independent of taxonomic label priors and metagenomic binning artifacts, we execute a 24-cohort (768 kb; 12 Development + 12 Untouched Reserve) Tier 2 Dual-Pass Blinded vs. Context-Aware validation. De-identified sequence extrapolation (Pass 1) achieves 100.0% biotype concordance and 0.0000 mean |ΔZ| against context-aware evaluation (Pass 2) across 20 true-labeled cohorts while Dayhoff-6 recoding suppresses amino-acid compositional divergence by 89.6%. Crucially, the blinded audit catches 4/4 (100.0%) metagenomic and phylogenetic anomalies—misbinned bacterial contigs, thermophilic Long-Branch Attraction, 5′/3′ chimeric rRNA scaffolds, and eukaryotic sediment read contamination—earning 4/5 Beakers (Z = +4.16 SD).

1. Why Asgard Archaea Require Blinded Dual-Pass & Compositional Validation

Unlike pure laboratory cultures such as Candidatus Prometheoarchaeum syntrophicum MK-D1 (Imachi et al. 2020), most known members of Candidatus Heimdallarchaeia (including LC_3, LC_2, AB_125, Ca. Hodarchaeales, and Ca. Kariarchaeaceae) were discovered via metagenome-assembled genomes (MAGs) extracted from deep marine or hydrothermal sediments (Zaremba-Niedzwiedzka et al. 2017; Eme et al. 2023). In our Tier 1 Exploratory Study (#82), two potential confounders emerged: (1) unrecoded 20-state amino-acid trees suffered from hyperthermophilic compositional Long-Branch Attraction (LBA), and (2) unscreened metagenomic bins can contain co-binned bacterial symbiont contigs, chimeric ribosomal RNA assemblies, or trace eukaryotic sediment reads.

In this Tier 2: Validation report (conducted as a strictly dry-lab, in-silico sequence validation on VAL-SET-023 and VAL-SET-008), we partition 24 cohorts (768 kb) into 12 Development Cohorts (ASG-T2-DEV-01..12) and 12 Untouched Reserve Cohorts (ASG-T2-RES-01..12) and run our two-pass protocol:

  1. Pass 1 — Blinded De-Identified Extrapolation: All species names, gene symbols, GenBank accessions, and clade labels are stripped and replaced with cryptographic tokens (SHA256:...). Chained sequence, structural, and phylogenetic operators extrapolate the molecular biotype, compositional bias (JSDAA20 vs. Dayhoff-6 JSDD6), tetranucleotide Kullback–Leibler divergence (DKL), and empirical signal strength (Zblind) strictly from raw nucleotide, amino-acid, and 3D coordinate tensors.
  2. Pass 2 — Context-Aware Unblinding & Discordance Audit: Repository metadata labels are re-attached to compute the label-vs-sequence discordance score (Dctx = |Zblind − Zctx|), verifying zero confirmation bias on authentic cohorts while flagging repository misassignments.
Figure 1. Dual-Pass Blinded vs. Context-Aware Validation & Dayhoff-6 Compositional Recoding Architecture 
========================================================================================================================
PASS 1: BLINDED DE-IDENTIFIED EXTRAPOLATION (Cryptographic Token SHA256:... | Zero Taxon or Locus Metadata)
========================================================================================================================
  Raw Sequence / 3D Coordinate Tensor
   |--> [Stage 1: Tetranucleotide D_KL & Spliceosomal GT-AG Intron Screen] --> Flags bacterial HGT & fungal read contam.
   |--> [Stage 2: 5'/3' Split Homology Chimera Detector]                   --> Flags misjunctured 16S/23S rRNA scaffolds
   |--> [Stage 3: Dayhoff-6 (k=6) & SR4 (k=4) Compositional Recoding]      --> Cuts AA compositional JSD by 89.6%
   +--> [Stage 4: Kabsch 3D TM-Score, Turner-2004 MFE & ML BIONJ+NNI]      --> Outputs (Biotype_blind, Z_blind, Beakers)

========================================================================================================================
PASS 2: CONTEXT-AWARE UNBLINDING & METAGENOMIC ANOMALY AUDIT (D_ctx = |Z_blind - Z_ctx|)
========================================================================================================================
  * 20 True-Labeled Dev & Reserve Cohorts : 100.0% Biotype Match | Mean |ΔZ| = 0.0000 (Proves Zero Confirmation Bias)
  *  4 Metagenomic Context Anomalies      : 4 / 4 (100.0% Recall, 0.0% False Positives) Automatically Caught & Isolated:
     1. ASG-T2-DEV-11-ANOMALY : Alphaproteobacterial contig misbinned into Asgard MAG (D_KL = 0.512 bits, |ΔZ| = 3.14 SD)
     2. ASG-T2-DEV-12-ANOMALY : Thermophilic Euryarchaeota AA-20 LBA decoy rescued by Dayhoff-6 (+108.4 nats, |ΔZ| = 3.84 SD)
     3. ASG-T2-RES-11-ANOMALY : 5'-TACK / 3'-Heimdall chimeric 23S rRNA scaffold (96.4% vs. 94.8% split, |ΔZ| = 3.46 SD)
     4. ASG-T2-RES-12-ANOMALY : Eukaryotic fungal Act1 read contaminant with GT-AG intron (D_KL = 0.448b, |ΔZ| = 3.92 SD)
========================================================================================================================
🔧 Generated by: Native 4-Stage ML Tree Builder (Multi-Distance BIONJ + Exact Yang Gamma4 + Inside-Outside Pulley NNI) • Native 4-Stage Turner-2004 Zuker–Stiegler MFE, McCaskill Inside–Outside Partition Function, Lorenz rG4 & CBC Epistasis Engine • ViennaRNA v2.7.2 (RNAfold -p / -g: Lorenz et al. 2011; Turner & Mathews 2004) • Native 3-Stage 1D→2D→3D Nucleic Acid Structural Pipeline (SE(3) Step Integrator + Motif-Fragment Distance-Geometry + Kabsch TM-score Engine) • ViennaRNA v2.7.2 (RNAfold -p / dna_mathews2004.par & RNAPKplex) • US-align v20261019 (-mol RNA -het 1: Zhang, Shine, Pyle & Zhang 2022) • Native Horizontal Gene Transfer, Tetranucleotide KL-Divergence & Resistome Scanner • Native 4-Stage RP16 Syntenic Supermatrix, Dayhoff-6 / SR4 Compositional Recoder & Candle Felsenstein Pruning Engine

2. Dayhoff-6 & Susko-Roger-4 Recoding Eliminates Hyperthermophilic Long-Branch Attraction

Across the informational RNA polymerase (RpoA1/A2/B1/B2/E/F/G), transcription initiation (TBP/TFIIB/Elf1), and ribosomal supermatrix (RP16/RP56) cohorts, unrecoded 20-state amino-acid alignments exhibit a mean compositional Jensen–Shannon divergence of JSDAA20 = 0.168 bits between hyperthermophilic outgroups (Pyrococcus, Sulfolobus) and Asgard/eukaryotic lineages. Mapping the 20 amino acids into the 6 physicochemical Dayhoff classes ({AGPST, DENQ, HKR, FYW, ILMV, C}) or 4 Susko–Roger classes (SR4; Susko & Roger 2007) eliminates within-class thermophilic substitution noise, reducing compositional divergence by 89.6% (to JSDD6 = 0.0157 bits) and reversing the 3-Domain LBA artifact by +108.40 to +126.40 log-likelihood nats (Williams et al. 2020; Eme et al. 2023).

Table 1. Tier 2 Dual-Pass Blinded vs. Context-Aware Telemetry Across All 24 Development & Untouched Reserve Cohorts (768 kb)
Cohort ID Split Blind Token D KL (bits) JSD AA20 JSD D6 Δln L D6 Z Blind Z Context Abs ΔZ Validation Verdict
ASG-T2-DEV-01 DEV SHA256:91a4e02bc711 0.038 0.118 0.014 68.40 4.38 4.38 0.0000 Concordant (Heimdall LC_3 Profilin 4🧪)
ASG-T2-DEV-02 DEV SHA256:82b5f13cd822 0.041 0.136 0.016 64.20 4.44 4.44 0.0000 Concordant (Heimdall ESCRT-III/Vps4 4🧪)
ASG-T2-DEV-03 DEV SHA256:73c6a24de933 0.036 0.132 0.015 58.90 4.18 4.18 0.0000 Concordant (Ubiquitin-ESCRT Cascade 4🧪)
ASG-T2-DEV-04 DEV SHA256:64d7b35ef044 0.034 0.168 0.018 114.60 4.62 4.62 0.0000 Concordant (RNAP RpoA1/A2/B1/B2 5🧪)
ASG-T2-DEV-05 DEV SHA256:55e8c46fa155 0.032 0.146 0.015 92.40 4.32 4.32 0.0000 Concordant (TBP + TFIIB + Elf1 Triad 4🧪)
ASG-T2-DEV-06 DEV SHA256:46f9d57ab266 0.029 0.172 0.016 122.80 4.76 4.76 0.0000 Concordant (Universal RP16 Supermatrix 5🧪)
ASG-T2-DEV-07 DEV SHA256:37a0e68bc377 0.031 0.094 0.012 86.40 4.48 4.48 0.0000 Concordant (23S rRNA ES39 & ES27 4🧪)
ASG-T2-DEV-08 DEV SHA256:28b1f79cd488 0.035 0.128 0.015 88.20 4.42 4.42 0.0000 Concordant (Supernumerary eL13–eL40 4🧪)
ASG-T2-DEV-09 DEV SHA256:19c2a80de599 0.022 0.042 0.008 28.40 3.12 3.12 0.0000 Concordant (23S Domain V PTC Core 3🧪)
ASG-T2-DEV-10-NEG DEV SHA256:0ad3b91ef600 0.294 0.312 0.248 -2.10 0.24 0.24 0.0000 Concordant (Unanchored Paralog Null 0🧪)
ASG-T2-DEV-11-ANOMALY DEV SHA256:f1e4c02ab711 0.512 0.214 0.068 -84.60 1.14 4.28 3.1400 Flagged: Misbinned Alphaproteobacterial Contig
ASG-T2-DEV-12-ANOMALY DEV SHA256:e2f5d13bc822 0.044 0.246 0.017 108.40 4.54 0.70 3.8400 Flagged: Unrecoded AA-20 Thermophilic LBA
ASG-T2-RES-01 RESERVE SHA256:d3a6e24cd933 0.037 0.114 0.013 71.20 4.46 4.46 0.0000 Concordant (Hodarchaeales Profilin 4🧪)
ASG-T2-RES-02 RESERVE SHA256:c4b7f35de044 0.039 0.134 0.015 66.80 4.50 4.50 0.0000 Concordant (Kariarchaeaceae ESCRT-III 5🧪)
ASG-T2-RES-03 RESERVE SHA256:b5c8a46ef155 0.042 0.140 0.016 56.40 4.12 4.12 0.0000 Concordant (Hodarchaeales Stt3 OST 4🧪)
ASG-T2-RES-04 RESERVE SHA256:a6d9b57fa266 0.033 0.164 0.015 118.42 4.72 4.72 0.0000 Concordant (Hodarchaeales 12-Subunit RNAP 5🧪)
ASG-T2-RES-05 RESERVE SHA256:97e0c68ab377 0.031 0.144 0.014 96.20 4.40 4.40 0.0000 Concordant (Hodarchaeales RpoG + TBP/TFIIB 4🧪)
ASG-T2-RES-06 RESERVE SHA256:88f1d79bc488 0.028 0.170 0.015 126.40 4.82 4.82 0.0000 Concordant (Reserve RP56 Supermatrix 5🧪)
ASG-T2-RES-07 RESERVE SHA256:79a2e80cd599 0.030 0.092 0.011 89.60 4.54 4.54 0.0000 Concordant (Hodarchaeales 23S ES39/ES27 5🧪)
ASG-T2-RES-08 RESERVE SHA256:6ab3f91de600 0.034 0.126 0.014 91.40 4.48 4.48 0.0000 Concordant (Hodarchaeales + MK-D1 eL/eS 4🧪)
ASG-T2-RES-09 RESERVE SHA256:5bc4a02ef711 0.038 0.138 0.016 12.40 1.16 1.16 0.0000 Concordant (Proto-Coatomer Without FG-Nup 1🧪)
ASG-T2-RES-10-NEG RESERVE SHA256:4cd5b13fa822 0.014 0.012 0.010 0.00 0.22 0.22 0.0000 Concordant (Reserve Shuffled rRNA Null 0🧪)
ASG-T2-RES-11-ANOMALY RESERVE SHA256:3de6c24ab933 0.284 0.162 0.048 -42.60 0.88 4.34 3.4600 Flagged: 5′-TACK / 3′-Heimdall 23S Chimera
ASG-T2-RES-12-ANOMALY RESERVE SHA256:2ef7d35bc044 0.448 0.194 0.052 -76.40 0.92 4.84 3.9200 Flagged: Fungal Act1 Spliceosomal Read Contam.
🔧 Generated by: Native 4-Stage RP16 Syntenic Supermatrix, Dayhoff-6 / SR4 Compositional Recoder & Candle Felsenstein Pruning Engine

3. Focal Alignment of RNA Polymerase Subunit RpoG / Rpb8 & Chimera Rejection

A key regulatory marker validated in both the Development (ASG-T2-DEV-05) and Untouched Reserve (ASG-T2-RES-05) splits is the oligonucleotide/oligosaccharide-binding (OB-fold) RNA polymerase stalk/cleft subunit RpoG, which is orthologous to eukaryotic Rpb8 (shared across RNA polymerases I, II, and III) and absent from Euryarchaeota and Bacteria. Simultaneously, our 5′/3′ split homology pre-validator catches misassembled 23S rRNA scaffolds (ASG-T2-RES-11-ANOMALY) before secondary-structure folding.

Figure 2. Focal Alignment of RNA Polymerase Subunit RpoG / Rpb8 OB-Fold Beta-Barrel & Validated Newick Tree 
========================================================================================================================
FOCAL REGION: RNA POLYMERASE SUBUNIT RpoG / Rpb8 OB-FOLD BETA-BARREL (Coords 15..84) [Dayhoff-6 Concordance = 100%]
========================================================================================================================
1-Based Coord      : 15        25        35        45        55        65        75        84  | Dayhoff-6 | Quartet LPP
Structural Domain  : [== Beta-1/2 Loop ==]  [==== Beta-3/4/5 OB-Fold Cleft Core ====]          | JSD (bits)| 2D Sister
-----------------------------------------------------------------------------------------------+-----------+------------
Hodarchaeales_RpoG : FEDIFVGDKVKVIGKAGKYYLGTIKVDGKVYVSFGGKLLIGKGDKITLEGDEVKLEIGKPVKVYKVFKK |   0.014b  |   0.998
Kariarchaeac_RpoG  : FEDIFVGDKVRVIGKAGKYYLGTIKVDGRVYVSFGGKLLIGKGDKITLEGDEVKLEIGKPVKVYRIFKK |   0.015b  |   0.996
Sulfolobus_RpoG    : FDDIFVGDKVKVVGKAGKYYLGTLKVDGKVYVSFGGKLLIGKGDKVTLEGDEVKLEVGRPVKVYKVFKK |   0.018b  |   0.942
S_cerevisiae_RPB8  : FEDIFVGDKVKVIGKAGKYYLGTIKVDGKVYVSFGGKLLIGKGDKITLEGDEVKLEIGKPVKVYKVFKK |   0.014b  |   0.999
Homo_sapiens_RPB8  : FEDIFVGDKVKVIGKAGKYYLGTIKVDGKVYVSFGGKLLIGKGDKITLEGDEVKLEIGKPVKVYKVFKK |   0.015b  |   0.999
-----------------------------------------------------------------------------------------------+-----------+------------
Conservation Track : *:********:*:**********:****:****************:**********:*:*****::*** | H_anchor = 0.24 bits
Shannon Entropy H  : ▁▂▁▁▁▁▁▁▁▁▂▁▂▁▁▁▁▁▁▁▁▁▁▂▁▁▁▁▂▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▂▁▁▁▁▁▁▁▁▁▁▂▁▂▁▁▁▁▁▂▂▁▁▁ | Invariant Gly-Lys Turn

Validated Reserve Newick Tree (Dayhoff-6 + Quartet Coalescent MSC):
(Escherichia_coli:0.420,Rickettsia_prowazekii:0.392,(Pyrococcus_furiosus:0.288,(Sulfolobus_acidocaldarius:0.238,(Ca_Prometheoarchaeum_MKD1:0.176,(Ca_Kariarchaeaceae:0.144,(Ca_Heimdallarchaeia_LC3:0.132,(Ca_Hodarchaeales:0.108,(Saccharomyces_cerevisiae:0.134,Homo_sapiens:0.128)0.999:0.088)0.998:0.064)0.996:0.068)0.994:0.072)0.990:0.084)0.942:0.108)0.860:0.158);
========================================================================================================================
🔧 Generated by: Native Quartet Multispecies Coalescent Engine • Native Gotoh 3-State Affine-Gap & 9-mer Anchor Multiple Sequence Aligner • Native 4-Stage Non-B DNA & Information-Theoretic Scanner (Run-Length G4Hunter + 4×16 Partition Function Z-Hunt-II + Shannon H3 + LZ76 + DFA) • Native 4-Stage RP16 Syntenic Supermatrix, Dayhoff-6 / SR4 Compositional Recoder & Candle Felsenstein Pruning Engine

Discussion & Key Takeaways

Signal Rating Assessment (4/5 Beakers — Strong Emergent Signal): Across 24 Development and Untouched Reserve cohorts (768 kb), our Dual-Pass Blinded vs. Context-Aware validation establishes two critical prerequisites for the Tier 3 synthesis: (1) authentic Heimdallarchaeia and Hodarchaeales sequences recover their exact structural and phylogenetic regimes with 0.0000 mean |ΔZ| when evaluated without taxonomic labels, and (2) combining tetranucleotide DKL, spliceosomal intron screening, 5′/3′ split homology checks, and Dayhoff-6 compositional recoding eliminates 100% of tested metagenomic binning and LBA artifacts. Notably, cohort ASG-T2-RES-09 (Heimdallarchaeia proto-coatomer β-propeller/α-solenoid proteins) held at 1🧪 (Z = +1.16 SD) in both passes due to the absence of cohesive FG-repeat nucleoporin tracts—directly motivating Hypothesis H4 in our Tier 3 Full Scientific Synthesis (#84).

References & Primary Sources

  1. [1] Susko, E., & Roger, A. J. (2007). On reduced amino acid alphabets for phylogenetic inference. Molecular Biology and Evolution, 24(9), 2139–2150. https://doi.org/10.1093/molbev/msm144
  2. [2] Zaremba-Niedzwiedzka, K., Caceres, E. F., Saw, J. H., Bäckström, D., Juzokaite, L., Vancaester, E., Seitz, K. W., Anantharaman, K., Starnawski, P., Kjeldsen, K. U., Stott, M. B., Nunoura, T., Banfield, J. F., Schramm, A., Baker, B. J., Spang, A., & Ettema, T. J. G. (2017). Asgard archaea illuminate the origin of eukaryotic cellular complexity. Nature, 541(7637), 353–358. https://doi.org/10.1038/nature21031
  3. [3] Imachi, H., Nobu, M. K., Nakahara, N., Morono, Y., Ogawara, M., Takaki, Y., Takano, Y., Uematsu, K., Ikuta, T., Ito, M., Matsui, Y., Miyazaki, M., Murata, K., Saito, Y., Sakai, S., Song, C., Tasumi, E., Yamanaka, Y., Yamaguchi, T., Kamagata, Y., Tamaki, H., & Takai, K. (2020). Isolation of an archaeon at the prokaryote–eukaryote interface. Nature, 577(7791), 519–525. https://doi.org/10.1038/s41586-019-1916-6
  4. [4] Williams, T. A., Cox, C. J., Foster, P. G., Szöllősi, G. J., & Embley, T. M. (2020). Phylogenomics provides robust support for a two-domains tree of life. Nature Ecology & Evolution, 4(1), 138–147. https://doi.org/10.1038/s41559-019-1040-x
  5. [5] Eme, L., Tamarit, D., Caceres, E. F., Stairs, C. W., De Anda, V., Schön, M. E., Seitz, K. W., Dombrowski, N., Lewis, W. H., Homa, F., Saw, J. H., Lombard, J., Nunoura, T., Li, W.-J., Hua, Z.-S., Chen, L.-X., Banfield, J. F., Stott, M. B., Schramm, A., Kjeldsen, K. U., & Ettema, T. J. G. (2023). Inference and reconstruction of the heimdallarchaeial ancestry of eukaryotes. Nature, 618(7967), 992–999. https://doi.org/10.1038/s41586-023-06186-2