This post uses this May 2025 study for selecting bacteria desired to be shifted.

Perturbations in gut microbiota in autism spectrum disorder: a systematic review [May 2025]

Bacteria	Direction
Eubacteriales, Klebsiella, and Clostridium	Statistically High
Oscillospira, Dorea, and Collinsella	Enriched
Streptococcus, Akkermansia, Coprococcus, and Dialiste	Depleted

Our approach is simple, we look up each on Microbiome Taxa R² Site. We then look at probiotics bacteria that are associated with desired shifts.

• Too High — rarely do probiotics reduce most bacteria
  • Eubacteriales [ order } – nothing found
  • Klebsiella [ genus } – nothing found
  • Clostridium [ genus } – nothing found
  • Oscillospira – nothing found
  • Dorea [ genus } – nothing found
  • Collinsella [ genus } – nothing found
• Too Low
  • Streptococcus [ genus }
    • Streptococcus thermophilus
    • Lactococcus lactis
    • Lacticaseibacillus paracasei
    • Enterococcus faecalis
    • Bifidobacterium breve
    • Bifidobacterium catenulatum
  • Akkermansia [ genus }
    • Akkermansia muciniphila
    • Enterococcus faecium
    • Lactococcus cremoris
    • Bacteroides thetaiotaomicron
  • Coprococcus [ genus }
    • Enterococcus faecium
    • Lactococcus cremoris
    • Akkermansia muciniphila
    • Lactococcus lactis
    • Bacteroides thetaiotaomicron
    • Levilactobacillus brevis
    • Ligilactobacillus salivarius
    • Latilactobacillus sakei
    • Lactiplantibacillus plantarum
    • Bifidobacterium adolescentis
    • Bifidobacterium catenulatum
      

What is Microbiome Taxa R² Site

It based on associations determined from 1000 healthy individuals using shotgun analysis,

This is followed by a table of the bacteria associations. This table has probiotics identified (including several that are pending).

Bottom Line

Looking on the impact, the following three probiotics would be my top choice:

• Enterococcus faecium – Source
• Lactococcus cremoris – Source
• Akkermansia muciniphila – Source or Source

Why not just use studies? There is nothing wrong with using studies. Pages listing studies result are linked to from each of the above pages. The difference is that with this approach you get the relative impact (R_²) between probiotics which is not available from the studies. Also, some probiotics have very few studies, i.e. Lactococcus cremoris. Using both published studies and this tool gives the maximum coverage.