1) The almost universal association of petroleum with sedimentary rocks.
2) The close link between petroleum reservoirs and source rocks as shown by biomarkers (the source rocks contain the same organic markers as the petroleum, essentially chemically fingerprinting the two).
3) The consistent variation of biomarkers in petroleum in accordance with the history of life on earth (biomarkers indicative of land plants are found only in Devonian and younger rocks, that formed by marine plankton only in Neoproterozoic and younger rocks, the oldest oils containing only biomarkers of bacteria).
3) The close link between the biomarkers in source rock and depositional environment (source rocks containing biomarkers of land plants are found only in terrestrial and shallow marine sediments, those indicating marine conditions only in marine sediments, those from hypersaline lakes containing only bacterial biomarkers).
4) Progressive destruction of oil when heated to over 100 degrees (precluding formation and/or migration at high temperatures as implied by the abiogenic postulate).
5) The generation of petroleum from kerogen on heating in the laboratory (complete with biomarkers), as suggested by the biogenic theory.
6) The strong enrichment in C12 of petroleum indicative of biological fractionation (no inorganic process can cause anything like the fractionation of light carbon that is seen in petroleum).
7) The location of petroleum reservoirs down the hydraulic gradient from the source rocks in many cases (those which are not are in areas where there is clear evidence of post migration tectonism).
8 ) The almost complete absence of significant petroleum occurrences in igneous and metamorphic rocks (the rare exceptions discussed below).
The evidence usually cited in favour of abiogenic petroleum can all be better explained by the biogenic hypothesis e.g.:
9) Rare traces of cooked pyrobitumens in igneous rocks (better explained by reaction with organic rich country rocks, with which the pyrobitumens can usually be tied).
10) Rare traces of cooked pyrobitumens in metamorphic rocks (better explained by metamorphism of residual hydrocarbons in the protolith).
11) The very rare occurrence of small hydrocarbon accumulations in igneous or metamorphic rocks (in every case these are adjacent to organic rich sedimentary rocks to which the hydrocarbons can be tied via biomarkers).
12) The presence of undoubted mantle derived gases (such as He and some CO2) in some natural gas (there is no reason why gas accumulations must be all from one source, given that some petroleum fields are of mixed provenance it is inevitable that some mantle gas contamination of biogenic hydrocarbons will occur under some circumstances).
13) The presence of traces of hydrocarbons in deep wells in crystalline rock (these can be formed by a range of processes, including metamorphic synthesis by the fischer-tropsch reaction, or from residual organic matter as in 10).
14) Traces of hydrocarbon gases in magma volatiles (in most cases magmas ascend through sedimentary succession, any organic matter present will be thermally cracked and some will be incorporated into the volatile phase, some fischer-tropsch synthesis can also occur).
15) Traces of hydrocarbon gases at mid ocean ridges (such traces are not surprising given that the upper mantle has been contaminated with biogenic organic matter through several billion years of subduction, the answer to 14 may be applicable also).
link