doerlbh / random

here is a test
Xinxin is lovely
DNA computing? Seelig Nature
Multiplex CRISPR computing? GMC Nature
fitness evolution, ucsf, kortemme lab
David report first
Felix polymer into protein
and also chromatin
what a day
rDNA still worth pursuing but have to say more molecular and structural exploration would be much more interesting~
should not use random chosen states, but use the probablities calculated based on all thress sates
chromatin looping?
information into a review type thing
or it could be nothing
how to connect my dynamics rigid-body to DNA or protein looping
http://www.sciencemag.org/site/feature/data/prizes/ge/2010/fullwood.xhtml
another day of entering github journal…
MathBio PhD or SynBio or SysBio or CompBio???
Or Google Genomics?
FZ talk? Not gonna stop
Need to solidify some programs to apply
weekend plan
decide programs, send refer help
contact FY, sned email to several prf
visa
intel
internship sep
AZ, CO, CA ..
HPC as a fundamental capacity
HPC needs innovation: Efficiency, Reliability, Paralleism
Programmable Solutions Group
FPGA
IPG (Intel Processor Graphics)
CPU Cores
Markov Chain in other axis??? like in environment
Can we infer the entire enviorment based on a point
In a chessboard, how many grids do we need to mark down entire chess board.
stochastic gene expression
on JP paper, we can do stochastic modeling
thx jai & reply qi
for stochastic process of iR is there any better quantificaiton
iPSCs basals to think about in a simutaneous form
network transsion
7~8 hq, 8-11 422, 11-12 1em, 0-4 373, 4-5 1em, 5-7 531 bayesian in 3D, 7-8 1em, 8-10 mitapp… ,
catch up in ML, CV, AI, Stoch
filter
projects outline
labs outline
school ranks
end
resort ML AI projects and potential collaborations
book flights, compare schools
what do I wish to focus
make plan for courses next
adv physics, app math, sys bio
plan time
what makes a good biologidt
today I am learning Python. A new language that I used to think it is easy and not worthwhile.
It turned out to be a huge mistake.
Now I find it a very interesting programming tool that can enable me to do many things in data science and a geeky lifestyle.
I am very looking forward to it.!
review bicohem
back to bk lecture
PTM
Working with proteins
detecting protein by UV absorbance
spectrophotometer
copper-based assays
dye-based assays
protein purification
salting out
dialysis
centrifuge
chromatography
with UV detector and fraction collector
FPLC / HPLC
gel filtration
ion exchange chromatography
affinity chromatography
recombinant tags for affinity chromatography
e.g. GST, MBP, His, Strep
Rasmussen te al Nature 2011 Nobel protein
electrophoresis (SDS, Native, IEF)
gel matrix
SDS PAGE
silver staining 100 times more than blue staining
IsoElectric Focusing (IEF)
protein identifacation
2D PAGE = IEF + SDS PAGE
Western Blotting
one protein detection method
sandwich chamber
amino acid analysis
hydrolyzed sample?
N-terminal sequencing
Edman’s Method
protein cleavage
chemical cleavage
enzymatic cleavage
protein 3D structure determination
X-ray crystallography
homology modeling
SAXS SANS
NMR spectroscopy
EM
https://d11.baidupcs.com/file/63ede6346996a8e1ead9bd435c567ffb?bkt=p3-00008dd1247c3724f81d7bba83883f10f0b8&xcode=599367b6305f3ffe1b5eaa8cf80e39d6f80f1674f0ee75741682cb8519c2059f&fid=2802628340-250528-724911083285829&time=1509180020&sign=FDTAXGERLQBHSK-DCb740ccc5511e5e8fedcff06b081203-yohKRYCv2GhmpWduAWjjP%2Fllbx0%3D&to=d11&size=6953474048&sta_dx=6953474048&sta_cs=31699&sta_ft=iso&sta_ct=7&sta_mt=0&fm2=MH,Yangquan,Anywhere,,new_york,any&newver=1&newfm=1&secfm=1&flow_ver=3&pkey=00008dd1247c3724f81d7bba83883f10f0b8&sl=79364174&expires=8h&rt=pr&r=790002327&mlogid=6973958507398000983&vuk=2802628340&vbdid=1424611918&fin=R2016b_glnxa64_dvd1.iso&rtype=1&iv=0&dp-logid=6973958507398000983&dp-callid=0.1.1&hps=1&tsl=100&csl=100&csign=tpjPT045aQmUyt%2FnuZsuzqjdcy8%3D&so=0&ut=6&uter=4&serv=0&uc=713335145&ic=1552888790&ti=6c86a94138ec9ff8c88b804018b6e44e76c2c9ca120f0293&by=themis
oef
fractal learning
selective history
Biology of psychiatric diseases
Evolution, plenotypic evolution, protein evolution
Microbiome dynamics, ecology
metabolism of cancer (related to oxygen)

Naive Bayesian Intergration Approach

NETBAG

cortico-striatal-thalamic loops

cortico-striatal-thalamic loops

action seleciton
habit formation
selection/perception of important information
behavior control
adaptation in activity

Austism

deep layers 5/6
cortical neurons

MRI studies reveal underconnectivity

• structural and functional MRI show reductions in connectivity

Di Martino et al The Autism Brain Imaging Data Exchange, Mol. Psych 2014
Deshpande G et al. Identification. Frontier Human Neuroscience 2013

limbic system role in emotional intelligence
emotion and motivation
phan kl et al., functional neuroanatomy of emtion, neuroimage 2002
Roesch MR, Olson CR. Neuronal activity related to reward value and motivation in primate frontal cortex., sciecne 2014

Brain Expression Data (Allen Mouse Brain Atlas)
Lein et al. Genome-wide atlas of gene expression in the adult mouse brain. nature 2007
mouse ISH expression data
600 defined structures from standardized brain
20,000 genes in ~300,000 voxel

2015 version very well
voxel
brain connectivity data

Oh et al. Nature 2014. A mesoscale connectome of the mouse brain.
brain connectivity data (allen mouse brain connectivity atlas)

60k
maps in coordiate systems

bias with normal tissue

TRENDS in Neuroscience
Goal-directed output
Nacc neuron
DA-neuoron terminal

amygdala - emotional component
cortex prefrontoal - thinking and judging

nucleus occuleus

how brain process information

SICK dataset
SNLI dataset
Flickr30K

denotation graph
modeling denotational probabilities
multiple premise entailment
denotation of a sentence s is the set of possible worlds in which s is true - denotational semantics

define a small set of transformation rules, e.g. drop modifier, drop pps

hierarchical phrases

denotational vs. distributional similarities

distributional similarities
topically related but not necessarily in the same scene

semantic textual similarity

Semeval 2012 MSR Chen & Dolan
Modle DKPro similarity

textual entailment

conditional probability - PMI

Modeling Denotational Prbabilities

semantic embedding spaces
order embedding (vendrov et al, 2015)
binary

an embedding space for denoational probabilities

overlay

LSTM (Bowman et al., 2015)
eh=a8ojeh

gating individuals to selective population
population learning
gating in sc
2013 fusi mit dimensionality of neurons behaviors
demix selectivity
Rigotti et al 2013 Nature
fusi et al 2016
why grid cell doesn’t exist
place cells could be good representations
we shouldn’t resort to neurons that has simple representations
animals don’t need hippocampus to encode location
hippocampus should just be memory machanism
only sparse structure can reveal grid
but decoding from prefrontal cortex, not finding grid-like representations due to its lack of sparsity, but it is better at decoding the neural signals
can single neuron tell us anything?
the problem is the bias for the experimentalist
the point of view of human perception is different from single neurons
so many behaviors of the single neurons depends on local enviornments like receptive fields
the local properties are the innate property
coregistration of brains - transformation fields
test
jump NN
Thomas Parr - message passing
learning session pairing
Karl Friston paper 2018 Sci Rep

Genearlization gap - margin distributions
tal’s proposal
random label data
inductive bias

Margaret S Livingstone
talk
modules for specific domains
thompson 1980 thatcher img example
modules -> expertise
specialized circuitry -> fine distinctions
(experience)
clustering -> local connections -> expertise
expertise <-?-> modules/clusters
Functional domains: innate biases vs. experience
reason for innate biases: import regions, stereotypes, areas
reason for experience: text, faces
symbol training -> symbol patches
fixed locations!!!
any expertise level; any training order
-> experience-dependent segregation on a proto-map
face selectivity arises later in time
no domains at birth
experience produces domain
proto-org –experience–> category selective
or immature domain –maturation–> mature domain

using sleeping monkey data?
adult ventral visual hierarchy –> connectivity inferred from covariation
early retinopic map –> arcaro & livingstone, elife
2017
how does category arises?
category - eccentricity - curvature

domain <-?-> looking
looking precede domain formation
hand looking first? face looking is not innate
newborn looks at faces
ferrari et al 2006 movie
superior curicuous - glance
hafed & chen 2016
upper visual field biaes
experience is necessary
and sufficent
horton & hocking - odc prior
ODC - sorting strategry?
map is a universal principle of newborn cortex

bruce, desimone, gross 1981
evolve cell with GAN and neuron recording
will xiao

occuluded faces

implausible occulsion
body orientation tuning

charles XXX syndrome - lack of central vision - hullucinate faces