The future of genetic medicine

Question 1

Splenic mass

Answer 1

2/3 hemangiosarcoma which have poor life expectancy
1/3 = benign

can’t know unless you do expensive sx to take it out. Half of dogs end up euthanized because ppl dnt want to pay for sx if dog is going to die but could have ended up being benign in 1/3 of those cases

Question 2

how would you (in theory) test to see if hemangiosarcoma or a benign mass

Answer 2

1. Abdominocentesis (run fluid in test to determine if mass is hemangiosarcoma or not)
2. Collect a blood sample (examine cell free DNA and epigenetic markers in blood sample)

Question 3

How would you make a test that runs abdominocentesis fluid to see if a mass is hemangiosarcoma or not

Answer 3

mRNA contains information about current state of cell (readout of complex cellular behavior the cell is currently doing; differences in how many rnas being expressed plays a role in cell going from benign to malignant so we would have to find genes that distinguish hemangiosarcomas from other dxs

Question 4

Assays you can look at for functional genomic tools

Answer 4

• can measure amount of TF/ Histones
  GRO- seq: measures amount of RNA polymerase (Pol II), measures the quantity of stable mRNA, measures translational rates; look at mRNA, RNA, translation ect

Question 5

epigenetic markers

Answer 5

look for epigenetic markers by looking for what our genome is doing; assays give quantities of a particular mark for each position in genome

Question 6

how would you divide patients into groups

Answer 6

based on molecular signals then goal is to use groups to see individual types of diagnoses (leverage molecular profile to get information about differential diagnoses); once you do this compare with clinical data

Question 7

gene expression clustering

Answer 7

use heat maps to look at different types of genes and where they are most active or least active then compare clusters of gene types with clinical variables and see what type of therapy they respond best to

Question 8

breast cancer

Answer 8

• breast cancer has a 70 gene breast cancer gene signature and can use high and low risk tests to decide who to give chemo and hormonal therapy to (high risk) or who to only use hormonal therapy on (low risk)

Question 9

Transcription factors

Answer 9

cellular conductors (these are the leaders); if you can find how cells are coordinating behaviors easier to see what behaviors are and figure out how to control them

Question 10

How do transcpriton factors work

Answer 10

transcription factor that recognizes certain DNA sequence and recruits RNA pol and other things that make post translation modifications on chromatin then RNA pol sent into productive elongation over entire gene; this is all started by transcription factor

Question 11

Driving force for cancer

Answer 11

DNA sequence mutation which activate pathways which active transactors which turns on groups of genes which have functions like rapid growth or ability to move to another location; these kinds of activities= responsible for what malignant cells are doing hence they could be prognostic indicators

Question 12

do transcription factors drive groups of genes that predict poor clinical outcome?

Answer 12

idea is that transcription factors target group of genes that carry biological fx that is important in determining survival time of particular patient

Question 13

Target genes of transcription factors

Answer 13

ID binding site of transcription factors genome wide and connect it with genes it can target and find ones where expression of RNA where transfactor binding site correlates with expression of gene and those are targets

Question 14

ADM

Answer 14

• in subset of glioblastoma patients this is highly transcribed; target for 3 transcription factors that bind in locus near this gene
  ADM is correlated with survival regardless of subtype
  Low ADM survive longer
  so one transcription factor for ADM may play role in initiating gene expression program that -> poor survival

Question 15

transcription factors that correlate with clinical outcomes

Answer 15

they have found three
-CEBPB
- RELA
- RARG
together these form prognostic signature

Question 16

How to do a study of transcription factors to look for prognostic signatures

Answer 16

• enroll 100-200 dogs who have splenic mass
• remove mass
• determine whether sample is HSA or benign
• collect gene expression data from each sample
• ID signature of gene expression
• validate if this predicts for hemangiosarcoma or not and validate signature in independent cohort of dogs

Question 17

How to use blood sample to see if HGA in dog (in theory)

Answer 17

• in theory cancer will be harming tissues near by and cancer and dying tissues throwing off genetic material you can pick up in plasma; certain cancers examined show that you can detect cell free DNA from tumor in blood stream but other cancers that you can’t do that with (like gleoma bc blood brain barrier)

Question 18

cell free DNA

Answer 18

in plasma variety of degradation products of DNA in various stages of decay from dying cells (usually dying blood cells); DNA in plasma has 20 min clearance rate so this is v short lived and therefore good readout of what particular patient doing at any particular time

Question 19

donor derived DNA

Answer 19

• existing test for circulating cell free DNA
• if transplant fails donor organ starts throughing out DNA into plasma bc cells dying and can pick out donor vs recipient DNA

Question 20

fetal cfDNA

Answer 20

• also an existing test for circulating cell free DNA

- used to test for trisome

Question 21

how do you know DNA is from a certain type of cancer

Answer 21

look for somatic mutations (DNA sequence mutations you know are associated with a specific type of cancer); some gives you many more mutations than others

Question 22

detect non small lung cell carcinoma in human patients

Answer 22

strategy for using non cell DNA
look at driver mutations of different kinds (specifically NSCLC) and then do broader exon sequencing for cancers in general
minimize targets you have to look at while maximizing efficiency for finding cancer

Question 23

CAPSEQ

Answer 23

test developed to detect NSCLC from non small lung cell carcinoma tumors

Question 24

specificity

Answer 24

how freqnently when you call a patient as having cancer to they have cancer

Question 25

sensitivity

Answer 25

what fraction of patents that have cancer do you correctly call as having cancer

Question 26

perfect classifier

Answer 26

would have 100% sensitivity and 100% specificity (- specificity in axis) so line goes straight up and over

Question 27

DNA methylation

Answer 27

have taken antibodies and enriched for DNA that is methylated with in each one of these tumors; methylation is a post translational mark to DNA itself; when DNA methylated it correlated with lower activity of genes nearby (its an indirect measure of gene expression activity)

Question 28

pancreatic adenocarcinoma

Answer 28

-study looked at sex and age match for control looked at tumor resection and collecting parallel plasma and cell free DNA from plasma and taking plasma from sex and age match controls; sequence DNA to see where methylation sites in cell free DNA; there are regions where tumor gain and loose methylation and those are enough to classify PDACs from normal based on this study

Question 29

cfDNA methylation

Answer 29

accurately IDs tumors, this needs refinement before its used as a diagnostic test but it shows promise

Question 30

TSME

Answer 30

look at cancer markers in 3D sets; different tumors separate into different space in dimensionality reduction space