Pathway map

No proposed route runs from raw materials to self-copying RNA

We wrote four proposed routes as step tables, straight from the papers, and scored every hand-off with a rule we sealed before we started. A step counts as joined only if it can run on what the step before left behind, using a named physical process.

01The map

Each route, step by step

4routes scored
26steps
0 of 22hand-offs pass cleanly
27gaps, ranked
  • Yield published from the setting's own inputs
  • Yield published, but only from purified or bought inputs
  • No published step known to us
  • Published, but we could not open the full text
  • Hand-off breaks a hard rule
  • Hand-off leads to or from a missing step
  • No rule broken, but some checks lack numbers

No route runs from raw materials to a self-copying molecule

S1Small sugars

S2Nucleotides

S3RNA chains

S4Template copying

S5Sustained replication

Cyanosulfidic

10 steps · 8 of 9 hand-offs unexplained

S1 · Small sugars

Raw materials of the setting

Passes, not fully checkable

Sugars from cyanide, trappedC1 · S1

HCN -> glycolaldehyde and glyceraldehyde (trapped as cyanate oxazolidinones), Cu(I) photoredox

oxazolidinone 13 (from glycolaldehyde) ~15% on average per 4 HCN; 14 (from glyceraldehyde) 1-5%; ~30% of HCN lost to a black polymer

Yield from the setting's own inputs. Source

First break. F4 input source

2-thiooxazoleC2 · S1

glycolaldehyde + thiocyanic acid -> 2-thiooxazole

85% (isolated)

Yield only from purified inputs. Source

F4 input source

Pentose thionesC3 · S1

2-thiooxazole + glyceraldehyde -> pentose oxazolidinone thiones (1a)

up to 74% thione 1a in 24 h (70% ribo+arabino 1:1, 30% lyxo/xylo)

Yield only from purified inputs. Source

S2 · Nucleotides

Passes, not fully checkable

AminooxazolinesC4 · S2

thione 1a + cyanoacetylene + ammonia -> arabino- and ribo-aminooxazoline (10a)

15-23% 10a per pass from cyanovinyl thione 1b with NH4Cl 1 M at pH 8.5-10.5; 45% arabino + 33% ribo over two cyanovinylation/ammonolysis cycles at pH 10.5

Yield only from purified inputs. Source

Leads to or from a missing step

AnhydrocytidineC5 · S2

arabino-aminooxazoline + cyanoacetylene -> 2,2'-anhydrocytidine

published (Powner 2009) but full text paywalled; not read

Published, full text not read. Source

F3 solvent, F6 purified hand-off

Cytidine cyclic phosphateC6 · S2

anhydrocytidine + phosphate + urea in formamide -> cytidine-2',3'-cyclic phosphate

44% (co-phosphorylated with 8,2'-anhydro-adenosine, 34%); formamide, 100 C, 72 h

Yield only from purified inputs. Source

F3 solvent

Uridine by UV lightC7 · S2

UV: cytidine-2',3'-cyclic phosphate -> uridine-2',3'-cyclic phosphate (partial)

published (Powner 2009, Fig. 5) but full text paywalled; not read

Published, full text not read. Source

S3 · RNA chains

F4 input source, F6 purified hand-off

Short RNA chainsC8 · S3

wet-dry polymerisation of 2',3'-cyclic nucleotides (A, U, G, C) -> RNA oligomers

AUGC at pH 11: Y = 36% into oligomers; 6% 4-mer, 0.1% 10-mer (10 cycles)

Yield only from purified inputs. Source

S4 · Template copying

Leads to or from a missing step

Template copyingC9 · S4

template-directed copying of oligomers under the same conditions

no published step known to us

No published step.

S5 · Sustained replication

Leads to or from a missing step

Sustained replicationC10 · S5

sustained replication of a replicator made from these oligomers

no published step known to us

No published step.

Alkaline vent

6 steps · 5 of 5 hand-offs unexplained

S1 · Small sugars

Raw materials of the setting

Passes, not fully checkable

Formate and acetateV1 · S1

CO2 + H2 -> formate + acetate across Ni-FeS under pH and thermal gradients

formic acid 735.9 uM, acetic acid 11.15 uM after 18 h (ocean 6 C pH 5.7 / vent 75 C pH 11, Ni-FeS)

Yield from the setting's own inputs. Source

First break. Leads to or from a missing step

Small sugarsV2 · S1

formate/acetate -> C2-C3 sugars or aldehydes under vent conditions

no published step known to us

No published step.

S2 · Nucleotides

Leads to or from a missing step

NucleotidesV3 · S2

-> nucleobases and ribonucleotides under vent conditions

no published step known to us

No published step.

S3 · RNA chains

Leads to or from a missing step

RNA chainsV4 · S3

nucleotide polymerisation in vent pores

no published step known to us

No published step.

S4 · Template copying

Leads to or from a missing step

Template copyingV5 · S4

template copying

no published step known to us

No published step.

S5 · Sustained replication

Leads to or from a missing step

Sustained replicationV6 · S5

sustained replication

no published step known to us

No published step.

Warm pond

5 steps · 4 of 4 hand-offs unexplained

S1 · Small sugars

Not part of this route

S2 · Nucleotides

Raw materials of the setting

First break. F4 input source

NucleotidesP0 · S2

nucleotide synthesis in the pond

no published step known to us

No published step.

F6 purified hand-off

Cyclic nucleotidesP1 · S2

wet-dry cycling of 5'-AMP / 5'-UMP in hot-spring water -> cyclic mononucleotides, U dimers

abstract reports synthesis of cyclic mononucleotides and U dimers; no yield in abstract; full text not opened

Published, full text not read. Source

S3 · RNA chains

F4 input source, F6 purified hand-off

Short RNA chainsP2 · S3

wet-dry polymerisation of 2',3'-cyclic nucleotides (A, U, G, C) -> RNA oligomers

AUGC at pH 11: Y = 36%; 6% 4-mer, 0.1% 10-mer; G alone up to ~70% at pH 10

Yield only from purified inputs. Source

S4 · Template copying

Leads to or from a missing step

Template copyingP3 · S4

template-directed copying / ligation at pond pH

no published step known to us

No published step.

S5 · Sustained replication

Leads to or from a missing step

Sustained replicationP4 · S5

sustained replication

no published step known to us

No published step.

Ice and RNA

5 steps · 4 of 4 hand-offs unexplained

S1 · Small sugars

Not part of this route

S2 · Nucleotides

Raw materials of the setting

First break. F4 input source

Triplet building blocksI1 · S2

prebiotic synthesis of trinucleotide triphosphates (pppNNN)

no published step known to us

No published step.

S3 · RNA chains

Leads to or from a missing step

A 45-letter ribozymeI2 · S3

abiotic emergence of a 45-nt polymerase ribozyme (QT45)

no published step known to us

No published step.

S4 · Template copying

F4 input source, F6 purified hand-off

QT45 copies itselfI3 · S4

QT45 copies its complementary strand and itself from triplets in eutectic ice

(-) strand 0.24% and (+) strand 0.17% of template in 72 days; fidelity 94.1%/nt (64-triplet pool); 10.9% perfect full-length (-), 43.4% perfect (+) with defined triplets

Yield only from purified inputs. Source

F4 input source, F6 purified hand-off

Short RNAs multiplyI4 · S4

pH-freeze-thaw cycling: triplet-based exponential replication of short dsRNA by the TPR ribozyme

1.2-1.46-fold per cycle for short (+8 to +14 nt) products; 29-nt ribozyme fragment up to 1.26-fold per cycle; A_D duplex 9x (+) and 31x (-) over 21 cycles

Yield only from purified inputs. Source

S5 · Sustained replication

Leads to or from a missing step

Sustained replicationI5 · S5

exponential replication of a whole polymerase ribozyme by itself

no published step known to us

No published step.

Real data. Read each column top to bottom. Open a step for its full name, yield and source. Hand-offs are scored with the sealed compatibility rule (sha256 bfa572ed...). Source: experiments/pathway_map/map.json

02First breaks

Two routes break before their first step, two right after it

  1. Cyanosulfidic

    Breaks after step 1

    The sugars come out trapped as oxazolidinones (Ritson 2012), and the next step needs free glycolaldehyde. It still gets furthest, to short RNA chains, but 8 of its 9 hand-offs are unexplained.

  2. Alkaline vent

    Breaks after step 1

    Formate and acetate form across iron-sulfide walls (Altair 2025). No vent step to sugars or nucleotides has been published.

  3. Warm pond

    Breaks before step 1

    The pond has no nucleotide source. Both pond papers start from laboratory nucleotides.

  4. Ice and RNA

    Breaks before step 1

    Nothing makes the triplet building blocks, and every RNA input, the ribozyme included, is supplied purified.

The alkaline vent ranks first by the rule, with no hard fail. That is an artefact: 5 of its 6 steps have no published chemistry, so there is nothing to break.

03Top gaps

Sustained replication is missing from every route

No route has any published step in which a replicator makes more copies of itself than it loses. The only template copying with a yield uses a purified, selected ribozyme with purified triplets, primers and a hexamer.

4 of 5 stages block all four routes

  1. 1-4Sustained replication of a whole replicator S5

    Routes blocked4 of 4
    Routes with any yield0 of 4
  2. 5-15Nucleotides from the setting's own raw materials S2

    Routes blocked4 of 4
    Routes with any yield1 of 4
  3. 16-20Template copying S4

    Routes blocked4 of 4
    Routes with any yield1 of 4
  4. 21-24RNA chains with purines, without bought nucleotides S3

    Routes blocked4 of 4
    Routes with any yield2 of 4
  5. 25-27Free small sugars from one-carbon chemistry S1

    Routes blocked2 of 4
    Routes with any yield2 of 4
Real data. 27 gaps, ranked: first those no route has bridged even with purified inputs, then by routes blocked. Bars run from zero to all four routes. Source: experiments/pathway_map/map.json (gaps)

04The rule

Nine hard fails decide whether a hand-off breaks

Step k+1 may use what step k leaves only through named physical processes from a closed list of 14, such as drying, heating, mixing or thermophoresis. The rule was hashed before any route was encoded, and it caught all 28 planted test cases first.

Hard fails

  • F1 pH
  • F2 temperature
  • F3 solvent
  • F4 input source
  • F5 concentration
  • F6 purified hand-off
  • F7 process validity
  • F8 redox
  • F9 setting

Warnings

  • W1 untested interference
  • W2 thin concentration
  • W3 claimed separation
Every hand-off, with the reason
RouteHand-offVerdictFlagsWhy
Cyanosulfidicentry->C1Passes, not fully checkablenoneFeedstocks of the scenario.
CyanosulfidicC1->C2Breaks the ruleF4Glycolaldehyde would have to be released from its oxazolidinone and the pot acidified; no named process does either.
CyanosulfidicC2->C3Breaks the ruleF4Most favourable reading: 2-thiooxazole crystallises from water (Stairs 2017), the crystals are redissolved by rain (acid left in the mother liquor; pH 6 stated by us, inside the rule's [0, 7] bound), concentrated 6x and warmed. Glyceraldehyde is still not available as a free sugar.
CyanosulfidicC3->C4Passes, not fully checkablenoneStairs 2017 crystallised all thione diastereomers (a lab crystallisation, counted as demonstrated; red-team target). NH3 uptake (gas exchange) raises pH by 2 units to 9 (the rule's cap).
CyanosulfidicC4->C5Leads to or from a missing stepW1Phosphate-buffered pH < 7 is needed; CO2 uptake (gas exchange) lowers pH by 2 units to 7.
CyanosulfidicC5->C6Breaks the ruleF3 F6Drying onto evaporite and heating; the source step runs in formamide, which no listed process supplies.
CyanosulfidicC6->C7Breaks the ruleF3 W1Back to the pool for UV; formamide has to be removed, which no listed process does.
CyanosulfidicC7->C8Breaks the ruleF4 F6 W1Same pool; purines (cAMP, cGMP) are not made by this route.
CyanosulfidicC8->C9Leads to or from a missing stepW1Into a missing step.
CyanosulfidicC9->C10Leads to or from a missing stepnoneInto a missing step.
Alkaline vententry->V1Passes, not fully checkablenoneFeedstocks of the scenario.
Alkaline ventV1->V2Leads to or from a missing stepnoneThermophoretic enrichment in rock cracks, at the rule's 1000x cap (Matreux 2024).
Alkaline ventV2->V3Leads to or from a missing stepnoneBetween missing steps.
Alkaline ventV3->V4Leads to or from a missing stepnoneBetween missing steps.
Alkaline ventV4->V5Leads to or from a missing stepnoneBetween missing steps.
Alkaline ventV5->V6Leads to or from a missing stepnoneBetween missing steps.
Warm pondentry->P0Breaks the ruleF4Feedstocks of the scenario.
Warm pondP0->P1Breaks the ruleF6From a missing step.
Warm pondP1->P2Breaks the ruleF4 F6Same pond. G and C cyclic nucleotides are not made by P1 (only A and U tested); P1's conditions are unread.
Warm pondP2->P3Leads to or from a missing stepnoneInto a missing step.
Warm pondP3->P4Leads to or from a missing stepnoneInto a missing step.
Ice and RNAentry->I1Breaks the ruleF4Feedstocks of the scenario.
Ice and RNAI1->I2Leads to or from a missing stepnoneBetween missing steps.
Ice and RNAI2->I3Breaks the ruleF4 F6From a missing step.
Ice and RNAI3->I4Breaks the ruleF4 F6 W1Same ice; Attwater's replicator is a different ribozyme with primers and a template supplied.
Ice and RNAI4->I5Leads to or from a missing stepnoneInto a missing step.

05Literature view

The literature runs further, to compartments and evolution, but with gaps at every stage

These statuses come from the literature and cover the whole path to evolution. The scored map above covers the first five stages, under the compatibility rule hashed in advance.

StepBest evidenceStatus
Feedstocks (HCN, formaldehyde, phosphate, sulfides) in a stated settingSeveral settings proposed; each supplies a different subset.Partial
Amino acids and nucleobasesMade in many routes; 33 amino acids and all five nucleobases in Bennu (2025); all five nucleobases in Ryugu (2026).Met
Ribose and nucleotidesRibose found in Bennu, deoxyribose not (Furukawa et al. 2025); pyrimidine nucleotides (C and U only) made without free ribose (Powner et al. 2009).Partial
Activation (energy-carrying building blocks)Thioester routes in water at neutral pH (Singh et al. 2025).Partial
Polymers long enough to foldShown with purified inputs. Wet-dry cycling gives short mixed RNA chains without activators, 0.1% reaching 10 letters (Caimi et al. 2025); continuity unscored.Partial
Template copyingQT45 copies itself and its complement in ice at about 0.2% yield in 72 days (2026).Partial
Exponential self-replicationShort RNA duplexes, including a ribozyme fragment, replicated exponentially from supplied triplets under freeze-thaw cycles (Attwater et al. 2025); never a whole replicator from scenario feedstocks.Not met
Compartments that grow and divide with their contentsPeptide sequence sets vesicle growth rate (Baba et al. 2026); membranes formed without a parent membrane, using enzymes (Khanal et al. 2026). Growth and division coupled to contents not shown.Partial
Darwinian evolutionRNA-catalysed evolution of a catalytic RNA (Papastavrou, Horning & Joyce, PNAS 2024), not open-ended.Not met

Scenarios

Alkaline hydrothermal vents

Explains a natural energy source (pH gradients) and CO2 reduction to formate and acetate.

Cannot yet anything beyond two-carbon products; nucleotides.

Surface ponds with wet-dry cycles

Explains concentration and joining of nucleotides by drying, without added activators.

Cannot yet chains stay short as breakdown balances growth; some natural spring waters destroy RNA.

Cyanosulfidic surface chemistry

Explains common origins for nucleotide, amino acid and lipid precursors under UV light.

Cannot yet its sugars come out trapped, and no named process frees them; no canonical purines.

Ice (eutectic freezing)

Explains concentrates RNA and slows breakdown; QT45 works here.

Cannot yet where the activated triplets come from.

06Files

The map is a data file anyone can rescore

Not checked: Powner 2009 and Šimonis 2026 were read as abstracts only (steps C5, C7 and P1 count as unread), and no supplementary files were read. An audit of the encoding by another agent is open (task #21). Results: experiments/pathway_map/RESULTS.md.