Defining neurotransmitter types

We systematically assigned neurotransmitter identity to each cell cluster based on the expression of canonical neurotransmitter transporter genes and synthesizing enzymes (Fig. ​(Fig.1e,1e, Fig. ​Fig.3,3, Extended Data Fig. ​Fig.3e,3e, Extended Data Fig. ​Fig.9,9, Supplementary Table 7). Criteria used are:

Glutamatergic (Glut): Slc17a6 (also known as Vglut2), Slc17a7 (Vglut1) or Slc17a8 (Vglut3).

GABAergic (GABA): (Slc32a1 (Vgat) or Slc18a2 (Vmat2)) and (Gad1, Gad2 or Aldh1a1).

Glycinergic (Glyc): Slc6a5.

Cholinergic (Chol): Slc18a3 (Vacht) and Chat.

Dopaminergic (Dopa): (Slc6a3 (Dat) or Slc18a2) and (Th and Ddc).

Serotonergic (Sero): (Slc6a4 (Sert) or Slc18a2) and (Tph2 and Ddc).

Noradrenergic (Nora): (Slc6a2 (Net) or Slc18a2) and Dbh.

Histaminergic (Hist): Slc18a2 and Hdc.

We used a stringent expression threshold of log₂(CPM) > 3 for these genes to assign neurotransmitter identity to each cluster.

These criteria are stringent as they require co-expression of both a neurotransmitter transporter and the corresponding key neurotransmitter synthesizing enzyme(s). They are also inclusive as alternative neurotransmitter synthesizing and releasing genes are included. For example, we included the vesicular monoamine transporter Slc18a2 (Vmat2) to all monoamine transmitters as well as GABA. It is known that in many midbrain dopamine neurons (in VTA and SNc), Aldh1a1 is used for synthesizing GABA in the absence of Gad1 or Gad2, and Slc18a2 is used for co-release of dopamine and GABA in the absence of Slc32a1⁵⁷.

Of note, a reported unconventional mechanism¹¹⁸ underlying co-transmission of dopamine and GABA by SNc dopaminergic neurons in the STR does not depend on cell-autonomous GABA synthesis but instead on presynaptic uptake from the extracellular space through the GABA transporter Slc6a1 (Gat1) while still relying on Slc18a2 for synaptic GABA release, which could make Aldh1a1 unnecessary in these cells. However, because Slc6a1 is widely expressed in all GABAergic neurons, as well as astrocytes, and even many subcortical glutamatergic neurons, it is unclear to us how widely applicable this unconventional mechanism (which bypasses all GABA-synthesizing enzymes) is. Therefore, we did not include Slc6a1 in our criteria in order to minimize false positives, even at the risk of having some false negatives.