A full genome sequence of a billion neurons by 2028
At the UCSF Decoding the Brain Academy last night, Dr. Tomasz Nowakowski plotted the progress of neurons sequenced over time, including the project planned for 2025. It has been following a Moore’s Law-like exponential curve for a decade now. So, we should have a sequence of every one of the 86 billion neurons in a human brain by 2032, and maybe all the glial cells too.
Why sequence every cell? The genetic makeup varies, as does the methylation. On the left, you can see clustering analysis of 121 different cell types in the thalamus alone. Out neurons are very different, and they vary over time. I was reminded of an amazing discovery Ed Boyden of MIT shared with me — they found HIV-like encodings being expressed in human neurons. This might be a transposon-like viral vector for horizontal gene transfer in the brain. This is a controversial finding, partially because it is difficult to detect the signal from the noise of the delta from the human reference genome, which is still Craig Venter. I asked Ed how many neuron types does he suspect we’ll find? “An infinite number. They are all different.”
When we asked Nowakowski about the compute capacity of a neuron, he lit up with excitement. By analogy to our machine learning neural nets with back prop, we now know that our neurons don’t just adjust weights at the synapse. Firing-feedback adjusts the weights in all of the ~1000 inputs embedded in the dendrite of each neuron. This is called synaptic tagging, and it serves as an overlay to the relaxation back to a nominal rate of firing, a process that is not fully understood.
From the other speakers on stage, left to right:
• Shawn Hervey-Jumper: in 50 years of neurosurgery advances, we have added just 4 months to expected lifespan
• Mercedes Paredes: in the pre-natal and infant brain (up to 6 months), neurons migrate great distances (1-10cm) by sending out an axon and pulling the neuron cell body along, a sequence of push and pulls. For a sense of scale, if the neuron was a car, it is like driving across America. How do they navigate over these distances?
• Christine Liu: We can keep brain tissues alive for weeks now (for glass probe insertion into neural cell bodies). Learning how to juggle can increase cortical thickness by 3%.
A full genome sequence of a billion neurons by 2028
At the UCSF Decoding the Brain Academy last night, Dr. Tomasz Nowakowski plotted the progress of neurons sequenced over time, including the project planned for 2025. It has been following a Moore’s Law-like exponential curve for a decade now. So, we should have a sequence of every one of the 86 billion neurons in a human brain by 2032, and maybe all the glial cells too.
Why sequence every cell? The genetic makeup varies, as does the methylation. On the left, you can see clustering analysis of 121 different cell types in the thalamus alone. Out neurons are very different, and they vary over time. I was reminded of an amazing discovery Ed Boyden of MIT shared with me — they found HIV-like encodings being expressed in human neurons. This might be a transposon-like viral vector for horizontal gene transfer in the brain. This is a controversial finding, partially because it is difficult to detect the signal from the noise of the delta from the human reference genome, which is still Craig Venter. I asked Ed how many neuron types does he suspect we’ll find? “An infinite number. They are all different.”
When we asked Nowakowski about the compute capacity of a neuron, he lit up with excitement. By analogy to our machine learning neural nets with back prop, we now know that our neurons don’t just adjust weights at the synapse. Firing-feedback adjusts the weights in all of the ~1000 inputs embedded in the dendrite of each neuron. This is called synaptic tagging, and it serves as an overlay to the relaxation back to a nominal rate of firing, a process that is not fully understood.
From the other speakers on stage, left to right:
• Shawn Hervey-Jumper: in 50 years of neurosurgery advances, we have added just 4 months to expected lifespan
• Mercedes Paredes: in the pre-natal and infant brain (up to 6 months), neurons migrate great distances (1-10cm) by sending out an axon and pulling the neuron cell body along, a sequence of push and pulls. For a sense of scale, if the neuron was a car, it is like driving across America. How do they navigate over these distances?
• Christine Liu: We can keep brain tissues alive for weeks now (for glass probe insertion into neural cell bodies). Learning how to juggle can increase cortical thickness by 3%.