>>> GOOD AFTERNOON EVERYONE.

I'M FRANCIS COLLINS.

IT'S A PLEASURE TO WELCOME YOU

TO THE WEDNESDAY AFTERNOON

LECTURE AND I ESPECIALLY WANT TO

WELCOME GUESTS WE HAVE WITH US

HERE IN THE FRONT ROWS WHO ARE

VISITING NIH.

STAND UP SO WE CAN SEE WHO YOU

ARE.

WELCOME TO THESE PARTS.

THEY ARE FINDING OUT WHAT WE DO

HERE.

[APPLAUSE]

THANK YOU VERY MUCH.

I THINK WE HAVE A REAL TREAT IN

STORE FOR US TODAY IN A LECTURE

FROM JONATHAN WEISSMAN WHO IS

OUR PRESENTER.

I THINK IN A CERTAIN WAY, I CAN

CLAIM JONATHAN AS A SIBLING BUT

YOU'LL HAVE TROUBLE AT FIRST

MANAGING WHY I WOULD SAY THAT.

IT'S BECAUSE HIS FATHER, SHERMAN

WEISSMAN WAS ALSO MY

POSTDOCTORAL MENTOR SO I THINK

THAT MAKES US SIBLINGS, DON'T

YOU THINK?

I MET JONATHAN LITERALLY WHEN HE

WAS PROBABLY 15 OR 16 IN THE

SUMMERTIME IN THAT LABORATORY AT

YALE.

JONATHAN WAS POKING AROUND WHILE

THE REST OF US WERE TRYING

DESPERATELY TO SEE IF WE COULD

LEARN SOMETHING THAT WOULD LEAD

TO A PUBLICATION AND I THINK IT

TURNED OUT REASONABLY WELL FOR

MANY OF THE PEOPLE THERE BUT

TURNED OUT EXCEPTIONALLY WELL

FOR JONATHAN.

HE BENT ON FROM THAT EXPERIENCE

AS A HIGH SCHOOL STUDENT TO

ENROLL AT HARVARD.

AFTER HARVARD, HE WENT ON TO GET

HIS Ph.D. IN PHYSICS AT MIT

WITH PETER KIM AND THEN OBTAINED

A POSTDOCTORAL FELLOWSHIP AND

TRAINED WITH ANOTHER GOOD FRIEND

OF MINE,.

SO JONATHAN SEEMS TO HAVE
CHOSEN

WISELY IN TERMS OF
LABORATORIES

TO TRAIN IN AND GOODNESS KNOWS,

HE BUILT UPON THAT NOW FOR A

REALLY REMARKABLE INDEPENDENT

CAREER.

HIS RESEARCH HAS FOCUSED IN

SEVERAL WAYS, CERTAINLY IN

CONSIDERABLE EFFORT IN

PRODUCTIVITY IN THE AREA OF HOW

PROTEINS FOLD INTO THEIR CURRENT

SHAPE AND HOW THAT IS DERANGED

IN THE PRESENCE OF DISEASE.

MOST RECENTLY, HOWEVER, JONATHAN

HAS BEEN INVOLVED IN APPLYING

VARIOUS TYPES OF HIGH-THROUGHPUT

TECHNOLOGIES TO DECODE GENOMES

AND TO UNDERSTAND HOW THEY WORK.

AND IF YOU ARE SOMEBODY WHO IS

PAYING ATTENTION TO THE

LITERATURE, YOU MIGHT HAVE SEEN

THE PAPER WHICH WAS JUST

PUBLISHED LAST FRIDAY IN CELL,

REALLY QUITE A REMARKABLE PAPER,

CALLED, RIBOSOME PROFILING OF

MOUSE EMBYRONIC STEM CELLS

REVEALS THE COMPLEXITY AND

DYNAMICS OF MAMMALIAN PROTEOMES.

ONE OF MY FORMER POSTDOCS

ALERTED ME TO THIS PAPER SENDING

IT ALONG WITH A PDF WITH A

ONE-SENTENCE COMMENT.

THE COMMENT WAS, JONATHAN

WEISSMAN ONLY HITS HOME RUNS.

THAT'S WHAT HE SAID.

SO PLEASE HEALTH CARE ME IN

WELCOMING A HOW MANY RUN HITTER

TO THE PODIUM WHO IS NOW AT THE

HOWARD HUGHES MEDICAL INSTITUTE

AND UCFF, JONATHAN WEISSMAN.

[APPLAUSE]

>> THANK YOU VERY MUCH FOR THE

KIND INVITATION AND

INTRODUCTION.

IT'S WONDERFUL TO BE HERE.

I DID PERCEIVE ONE THING, WHICH

IS, I CAME TO THE NIH AS A -- IN

A FAR EARLIER YEAR, 1956, THE

YEAR OF MY BIRTH AND MY FATHER

WAS HERE AND GOT HIS SCIENTIFIC

ART IS HERE AND I WAS CARRYING

HIM INTO THESE HALLOS MANY A

DAY, ALTHOUGH I CAN'T SAY I

REMEMBER EXACTLY WHICH ONES.

20 YEARS AGO OR SO, WHEN PEOPLE

STARTED TO TALK SERIOUSLY ABOUT

SEQUENCING HUMAN GENOMES,

ESTIMATES HAD BEEN THAT THE COST

SHOULD BE ABOUT 10 DOLLARS PER

NUCLEOTIDE AND SO WITH 3 BILLION

NUCLEOTIDES IN OUR GENOME IT

WOULD BE A 30 BILLION PROJECT.

SO ARGUEY, IF YOU THINK ABOUT IT

A MONUMENTAL HUMAN ACHIEVEMENT

THAN SENDING MAN TO THE MOON BUT

PRETTY EXPENSIVE.

10 YEARS OR SO LATER, THE COST

THROUGH A LOT OF CLEVER EFFORTS

HAD COME DOWN BECAUSE IT WAS

ABOUT A 300 MILLION DOLLAR

PROJECT.

AND THANKS TO EFFORTS LED BY

FRANCIS, PEOPLE STARTED TO TALK

ABOUT WHAT WOULD HAPPEN IF WE

COULD HAVE 1000 DOLLAR GENOME.

THAT WASN'T LED BY ANY

TECHNOLOGICAL INSIGHT ABOUT HOW

YOU DO IT, IT WAS LED IN LARGE

PART BY THE FACT THAT AN

EXPENSIVE MEDICAL TEST, MRI, WAS

ABOUT 1000 DOLLARS.

IF YOU COULD BRING THE COST OF

SEQUENCING DOWN TO 1000 DOLLARS,

YOU WOULD REALLY IN A SERIOUS

WAY, USHER IN THE ERA OF

PERSONAL GENOMICS.

AND OVER THAT TIME, MORE OR LESS

HAS BEEN ACHIEVED AND ONLY GOING

TO GET CHEAPER FROM HERE AND IN

2001, THE COST OF OR THE

PRODUCTION OF CARS UNDERGONE THE

SAME INCREASE AND EFFICIENCY

OVER THIS 10 YEARS.

A 40,000 DOLLAR BMW WOULD COST

ABOUT 25 CENTS.

BUT BE CAREFUL OF WHAT YOU WISH

FOR.

WHAT YOU GET OUT OF SEQUENCE,

YOU GET A LOT OF -- AND IN
MY

LAB, ALONG WITH MANY OTHERS HAVE

TAKEN UP THIS CHALLENGE AND

THOUGHT THAT THIS INCREDIBLE

TECHNOLOGIC ADVANCE OF BEING

ABLE TO SEQUENCE VIRTUALLY

ANYTHING CHEAPLY AND RAPIDLY,

ACCOUNTED NOW BE USED NOT JUST

TO DETERMINE THE SEQUENCE OF THE

GENOMES BUT TRY TO DECIPHER THE

INFORMATION THAT IS ENCODED

WITNESS THESE GENOMES.

WHAT I'M GOING TO TALK ABOUT

TODAY IS AN APPROACH FROM MY LAB

THAT LETS US LOOK AT THE PROCESS

OF TRANSLATION IN TERMS OF

RIBOSOME PROFILING.

WHAT I HOPE TO CONVINCE YOU OF

IS THAT THIS IS BOTH A POWERFUL

OHMICS TOOL FOR BROADLY LOOKING

AT TRANSLATION IN VIVO AND A

HIGH PRECISION TOOL FOR LOOKING

AT THE MECHANISM OF TRANSLATION

IN THE IN VIVO SETTING.

AND REALLY, AS A TRADE

MECHANISTIC BIOCHEMIST, I'M MORE

EXCITED IN MANY WAYS BY THE

LATTER POTENTIAL.

WHY SHOULD WE LOOK AT

TRANSLATION IN THE FIRST PLACE?

AFTER ALL, AS I SAID, THE WORLD

IS AWASH WITH LARGE DATASETS AND

THERE ARE MANY WHO WOULD ARGUE

WITH QUITE A BIT OF TRUTH BEHIND

IT THAT WHAT WE NEED IN BIOLOGY

NOW IS NOT MORE DATA BUT MORE

TIME IN THINKING ABOUT WHAT

THESE DATA MEAN.

I WOULD STILL ARGUE THERE IS A

CENTRAL PLACE FOR BEING ABLE TO

LOOK BETTER AT THE PROCESS OF

TRANSLATION AND AT THE MOST

BASIC LEVEL, THE REASON IS THAT

BY-AND-LARGE, GENES ARE ACTING

THROUGH THE PROTEIN PRODUCTS

THAT THEY ENCODE.

AND FOR EXAMPLE, RAPIDLY

GOING -- UP TO HALF OF THE

ENERGY OF THE CELLS IS GOING

INTO PROTEINS.

BUT EVEN BEYOND THAT, ONE CAN'T

USE MRNA AS A PROXY FOR PROTEIN

PRODUCTION, BECAUSE WE KNOW THAT

THERE ARE EXAMPLES OF

TRANSLATIONAL CONTROL AND IT'S

NOT POSSIBLE TO DEFINITIVELY

PREDICT PROTEIN PRODUCTS FROM

MRNA SEQUENCE.

SO IF SOMEONE IS GOING TO

ANNOTATE THE PROTEIN CODING

POTENTIAL OF THE GENOME, YOU

NEED EXPERIMENTAL INPUT AND THEN

FINALLY, ALTHOUGH THERE ARE MANY

BEAUTIFUL EXAMPLES OF

TRANSLATIONAL CONTROL, OUR BROAD

UNDERSTANDING OF THE USE OF

TRANSLATIONAL CONTROL AND THE

MECHANISMS OF TRANSLATIONAL

CONTROL FAR BEHIND THOSE OF

TRANSCRIPTIONAL CONTROL.

SO WE STEP BACK AND SAY, WHAT IS

EXPERIMENTS YOU LIKE TO DO?

IN THE MIND-SET OF THE,000

DOLLAR GENOME.

JUST LET YOUR IMAGINATION SAY,

WHAT WOULD I WANT TO DO IF I

COULD?

YOU LIKE TO BE ABLE TO LOOK

INSIDE THE CELL AND YOU'D LIKE

TO SEE WHERE EVERY RIBOSOME IS,

WHAT PROTEIN IT IS TRANSLATING,

HOW FAST IT IS MOVING, IS IT

STARTING?

STOPPING?

STALLING?

AND OF COURSE TO BE ABLE TO DO

THIS IN ANY SORT OF DIRECT

VISUALIZATION APPROACH IS

BROADLY GOING TO BE EXTREMELY

HARD.

AND IT'S NOT LIKELY TO HAPPEN IN

THE NEAR FUTURE.

THAT SAID, WHEN YOU FRAME THE

QUESTION THAT WAY, THERE IS A

PIECE OF INFORMATION THAT ACTS

AS A PROXY AND QUITE A GOOD

PROXY FOR MANY USES.

IT RELIES ON AN OLD OBSERVATION

FIRST DEVELOPED BY JOE PSYCHES

AND THEN IN IN A MORE COMPLETE

FORM BY WALTER.

THIS TAKES ADVANTAGE OF THE FACT

THAT A RIBE SOSOME WILL PROTECT

ABOUT 30 NUCLEOTIDES OF THE MRNA

MESSAGE FROM DIGESTION.

IF ONE TREATS EXTRACTS WITH

NUCLEASE,֜ AND EXTRACTS
PROTECTED

MRNAs AND THEN SEQUENCES THEM,

YOU GET AN EXACT HISTORY OF

WHERE THE RIBOSOME WAS AT THE

TIME IN WHICH THE CELL WAS

DISRUPTED.

WHAT CHANGED OVER THE LAST 40

YEARS OVER SO IS OUR ABILITY TO

SEQUENCE THESE SMALL MRNAs.

SO NOW IN LUMINA SEQUENCE WE GET

HUNDREDS OF MILLIONS OR BILLIONS

OF THESE DIFFERENT SEQUENCES.

AND THAT'S A LARGE NUMBER BUT

IT'S REALLY A LARGE NUMBER AND

ULTIMATELY A PROBLEM.

EVEN IN A COMPLEX GENOME, THERE

ARE ONLY ABOUT 15 MILLION

DIFFERENT PLACES THAT A RIBOSOME

COULD BE.

SO IF YOU COULD GET HUNDREDS OF

MILLIONS, YOU CAN BUILDUP A

HISTOGRAM OF WHERE EVERY

RIBOSOME IS IN THE CELL AND HOW

OFTEN THE POTENTIAL POSITION IS.

SO STEPPING BACK THE TYPE OF

INFORMATION THAT YOU CAN GET IS,

YOU CAN IMAGINE IF YOU HAD TWO

MESSAGE THAT IS ARE EQUALLY WELL

TRANSCRIBED, SO SAME LEVELS OF

THESE MESSAGE BUT ONE IS WELL

TRANSLATED AND OTHERS POORLY

TRANSLATED AND THIS WOULD BE

REFLECTED IN THE DIFFERENCE OF

THE NUMBER OF RIBOSOME

PROTECTIVE FOOTPRINTS REFLECTING

THE NUMBER OF RIBOSE ON THESE

MESSAGES.

YOU CAN DISTINGUISH WHETHER

LOOKING AT A SUPER HIGH WHERE

THE RIBOSE ARE STREAMING DOWN

THE MESSAGE AND ONES WHERE YOU

HAVE TRAFFIC JAMS.

WITH TRAFFIC JAMS BUNCHING UP OF

RIBOSE REFLECT THE BUNCHING UP

OF THESE RIBOSE.

A BIT MORE DETAIL THE WAY A

TYPICAL EXPRESSION EXPERIMENT

WOULD GO WOULD BE TO DIVIDE THE

CELL EXTRACT IF TWO AND ONE YOU

CREATE THESE RIBOSOME PROTECTIVE

FRAGMENTS AND ANOTHER YOU JUST

RANDOMLY FRAGMENT THE MRNA AND

CONVERT THESE TO DNA AND COUNT

THE POSITION AND THE NUMBER OF

THESE DIFFERENT FRAGMENTS BY

DEEP SEQUENCING AND ON THIS

SIDE, FROM THE RANDOMLY

FRAGMENTED MRNA, YOU GET

INFORMATION ABOUT THE MRNA

ABUNDANCE AND THE SEQ EXPERIMENT

QUALITATIVELY THE SAME TYPE OF

INFORMATION YOU WOULD HAVE

GOTTEN FROM THE PIKE ROW ARRAY

TYPE OF EXPRESSION STUDY.

BUT ON THIS SIDE, BY COUNTING

THE PLACE AND NUMBER OF RIBOSOME

FOOTPRINTS ON EACH MESSAGE, YOU

GET A DIRECT MEASURE OF THE RATE

AT WHICH THAT MESSAGE IS BEING

TRANSLATED INTO POLYPEPTIDES.

AND HERE IS ACTUALLY SOME

GENOME-WIDE DATA FROM OUR FIRST

EXPERIMENT THAT IS FROM -- EACH

OF THESE POINTS IS A GENE.

THESE ARE FULL BIOLOGICAL

REPLICATES MEASURING MRNA

ABUNDANCE OR RATES OF

TRANSLATION AND THE BOTTOM LINE

IS WE CAN NOW MEASURE

TRANSLATION RATES, THE RATE OF

PRODUCTION OF PROTEINS FROM EACH

OF THE GENES WITH THE SAME

SPEED, DEPTH, COST, ACCURACY, AS

WE COULD WITH THE BEST MRNA.

FROM THIS INFORMATION ROUGHLY

SPEAKING, EQUALLY ACCESSIBLE AND

THE QUESTION IS, DID WE GAIN

MORE INFORMATION, MORE

INFORMATION ABOUT HOW MUCH THESE

PROTEINS ARE BEING MADE OR IS IT

A FANCY WAY OF MEASURING MRNA

ABUNDANCE?

THE ANSWER IS, THERE IS QUITE A

BIT OF INFORMATION AT THE

TRANSLATION LEVEL AND I'LL SHOW

YOU ONE SLIDE THAT I THINK

NICELY ILLUSTRATES IT.

I'M SWITCHING NOW FROM A YEAST

EXPERIMENT TO A BACTERIAL

EXPERIMENT AS YOU'LL SEE A

DETAIL OF HOW BACTERIA

TRANSLATES MESSAGES FOR

ILLUSTRATIVE PURPOSES.

AGAIN, THESE ARE E.COLI DATA AND

BIOLOGICAL REPLICATES FOR EACH

OF THE 4000 OR SO DIFFERENT

GENES IN E.COLI.

VERY HIGH PRECISION

MEASUREMENTS.

ABOUT 5 ORDERS OF MAGNITUDE

DIFFERENCES IN THE DIFFERENT

EXPRESSION LEVELS OF THE

DIFFERENT MESSAGES.

AND HERE IS WHY BACTERIA WERE

NICER ILLUSTRATING TRANSLATIONAL

CONTROL WHICH IS THAT MANY

GENES -- ON THE SAME PHYSICAL

PIECE OF MRNA, THEY'LL BE TWO

OPEN READING FRAMES, EACH

INDEPENDENTLY TRANSLATED.

AND WE CAN ASK HOW WELL IS THE

RATE OF TRANSLATION THEN

CORRELATED WITH THE GENE B IN

THE SAME MESSAGE?

ALTHOUGH THERE IS A VERY STRONG

STATISTICAL CORRELATION, IT

WOULD NEVER HAPPEN BY CHANCE AND

THAT THE CORRELATION COEFFICIENT

IS ONLY ABOUT .115.

SO ABOUT 88% OF THE VARIANTS IS

DIFFERENT WITH HUGE DIFFERENCES

WHICH DIFFERENT PROTEINS ARE

BEING EXPRESSED IS SIMPLY

INVISIBLE TO A MEASUREMENT OF

MRNA ABUNDANCE.

IF YOU WERE TO MEASURE MRNA,

THEY WOULD BE EXPRESSED IN THE

SAME LEVEL AND YOU WOULD BE

MISSING THIS INFORMATION.

AND WE HAVE NOW DONE THIS IN A

FULL -- SO THOSE WERE EXAMPLES

AS REALLY MORE FULL BIOLOGICAL

EXAMPLES OF TERRIFIC

POSTDOCTORAL FELLOW IN MY LAB,

HAS LOOKED AT A VERY DETAILS

IN-DEPTH EXPRESSION ANALYSIS IN

THE LEVEL OF TRANSLATION THROUGH

THE MYOTICS PROGRAM IN MICE.

AND HONESTLY THIS IS MORE OF AN

ADVERTISEMENT THAN IT IS OF

ACTUAL DESCRIPTION OF THE

EXPERIMENT.

JUST TO SAY THAT YOU CAN REALLY

DO THIS AS A FULL-SCALE GENE

EXPRESSION EXPERIMENT AND YOU

GET PRECISE PRECISION ABOUT

EXPRESSION AND PARTICULARLY SO

REMARKABLE ANALYSIS THAT THESE

ARE THE TIME POINTS IN MYOSIS IN

EACH OF THE GENES AND IT NOW

BECOMES STANDARD CLUSTERING TYPE

STRATEGY WHERE YOU PUT TOGETHER

GENES OF THE SAME TIME.

SHE WAS IN ONE FAIL SWOOP ABLE

TO FIND ALL OF THE GENES THAT

ARE INVOLVED IN CENTRAL PROSAND

RECOMBINATION AND NOT THE MOST

COMPLEX FORMATION.

IT'S VERY EXTENSIVELY STUDIED

CENTRALLY DURING THE MYOSIS AND

IN ADDITION TO THIS WAS ABLE TO

IDENTIFY A NUMBER OF NOVEL NEW

PLAYERS AND NEW PLAYERS IN THESE

PROCESSES.

SO AGAIN, BOTTOM LINE OF THIS IS

THAT I THINK RIBOSOME PROFILING

AS AN EXPRESSION TOOL IS NOW A

FAIRLY MATURE TECHNOLOGY.

BUT BEYOND COMING UP TO THE

LEVEL OF EXPRESSION OF EACH

DIFFERENT PROTEIN, YOU OF COURSE

BY ITS NATURE GET POSITION

SPECIFIC INFORMATION AND START

TO DEFINE EXACTLY THE MESSAGES

BEING TURNED INTO PROTEINS.

SO IT'S AN EXAMPLE.

YOU CAN SCHEMATICALLY, IF THERE

WAS AN OPEN READING FRAME, THAT

ENCOMPASSED ONLY PART OF THE

MESSAGE, YOU WOULD EXPECT THE

MRNA FRAGMENTS TO COVER THE

ENTIRE MESSAGE WHEREAS THE

RIBOSOME PROTECTIVE FRAGMENTS

WOULD GO ON UNTIL THEY STOP AT

THE STOP.

IN THIS WAY, YOU ARE DOING A

PROTEOMICS EXPERIMENT DEFINING

THE HIGH DEFPROTEIN THAT IS

BEING ENCODED.

YOU CAN SEE THIS VERY NICELY IN

THE METAGENE ANALYSIS WHERE YOU

CAN LINE UP ALL THE GENES FROM

THE START OR THE STOP, YOU SEE

THAT THE RIBOSOME֜ PROTECTIVE

FRAGMENTS START UP STREAM OF THE

START AND THEN JUST UP STREAM OF

THE STOP.

YOU ALSO SEE THERE IS A VERY

NICE PERIOD OF DATA AND THAT

REFLECTS THE SIMPLE FACT THAT OF

COURSE THE RIBOSOME DOESN'T

SMOOTHLY GO ALONG BUT DECODES A

CODON AND GOES ON TO THE NEXT

ONE.

AND YOU CAN SEE THAT IN A GLOBAL

ANALYSIS VERY NICE FRAMING

INFORMATION FROM THE RIBOSOME

CONNECTED FRAGMENT.

AND THAT'S IMPORTANT FOR TWO

REASONS.

ONE, IT'S REALLY PRETTY

DEFINITIVE EVIDENCE THAT WE ARE

LOOKING AT TRANSLATION BECAUSE

IT WOULD BE VERY HARD TO

UNDERSTAND HOW YOU WOULD GET

THIS OTHERWISE, BUT MORE

IMPORTANTLY, IT MEANS WE ARE NOT

JUST SAYING MORE OR LESS WHAT

WEDGEONS ARE BEING TRANSLATED.

WE ARE SAYING THE EXACT IS

TRANSLATED SO WHERE

FRAMESHIFTING OR REGIONS BEING

READ IN TWO FRAMES, WE CAN

DIRECTLY EXTRACT THAT FROM THE

DATA.

THE TRANSLATION ITSELF OF COURSE

IS A DYNAMIC PROCESS.

WHAT I HAVE BEEN SHOWING YOU WAS

SNAPSHOTS OF WHERE RIBOSE WERE

AT A DEFINED TIME WHEN WE OPENED

UP THE CELL.

AND BY THE USE OF SPECIFIC

INHIBITORS IN THE PROCESS OF

TRANSLATION, WE CAN START

NOTHING OTHER THAN A TIME SERIES

OF SNAPSHOTS.

AND THE INHIBITORS WE USE FOR

FIRST ONE CALLED -- A VERY

SPECIFIC INHIBITOR OF THE

PROCESS OF TRANSLATION

INITIATION BUT DOESN'T INHIBIT

THE ONGOING ELONGATION OF

RIBOSOME AND THE SECOND THAT MAY

BE MORE FAMILIAR TO MANY OF YOU

IS INHIBITOR OF TRANSLATION

ELONGATION.

SO WE'LL STOP RAPIDLY RIBOSE IN

THEIR TRACKS REGARDLESS OF WHERE

THEY ARE IN THIS PROCESS.

NOW, BY TREATING CELLS WAITING

FOR DIFFERENT TIME PERIODS AND

THEN QUENCHING THE ELONGATION,

ONE CAN CHASE THE RIBOSE FROM

THE BEGINNING TO THE END OF THE

MESSAGE AND BY FOLLOWING THIS

LOOK AT THE KINETICS OF

TRANSLATION.

SO AGAIN, YOU CAN SEE THIS IN

THE METAGENE ANALYSIS WHERE IF

YOU DON'T TREAT CELLS, YOU SEE

DENSITY OF RIBOSE ACROSS

MESSAGES ON AVERAGE AND THEN AS

YOU TREAT FOR INCREASING TIMES,

YOU SEE SPECIFIC DEPLETION AND

YOU SEE THIS FRONT OF

TRANSLATION MOVING FORWARD AT A

STEADY PACE AND ONE CAN PROCESS

AND THEN SEE ON AVERAGE THE

TRANSLATION IS PERCEIVING AT A 5

1/2 AMINO ACIDS PER SECOND WHICH

IS COMFORTING TO US BECAUSE

CLASSIC STUDIES USE THE STRATEGY

LOOKING AT A COUPLE OF MESSAGES

GOT IT MORE OR LESS THE SAME

NUMBER.

AND THAT REALLY SAID THE PROCESS

IS WORKING AND THAT HERRING TONE

IN WAS INTERFERING WITH

ELONGATION.

SO THIS SETS US UP FOR THE

REASON WE DID THIS EXPERIMENT.

IS THIS ELONGATION ON AVERAGE,

IS IT AN IMPORTANT MECHANISM FOR

CONTROLLING SYNTHESIS?

SO ARE CERTAIN CLASS OF PROTEINS

ELONGATED AT ONE RATE AND OTHER

CLASSES AT A DIFFERENT RATE?

FROM THIS WE GOT A SPECTACULAR

NEGATIVE RESULT.

WE PARSED UP THE RIBOSOMES BY

WHETHER THEY HAD GOOD CODON

USAGE OR POOR CODON USAGE,

HIGHLY EXPRESSED, LONG OR SHORT

OR SECRETED OR SOLUBLE AND FOR

ALL OF THEM WE GOT 5 1/2 AMINO

ACIDS PER SECOND.

THAT DOESN'T MEAN THERE AREN'T

SOME SMALL SUB SETS OF RIBOSE OF

MESSAGES FOR ELONGSIGATION IS

REGULATED.

UNDOUBTEDLY THERE WILLv: BE.

BUT IT DOESN'T MEAN UNDER

DIFFERENT CONDITIONS THE RATE OF

ELONGATION MIGHT NOT BE

REGULATED AND I THINK WE NOW

HAVE THE TOOLS TO INVESTIGATE

THOSE TWO QUESTIONS BUT IT MEANS

ON AVERAGE WE HAVE THIS ONE

UNIVERSAL THAT TELLS US THE RATE

AT WHICH TRANSLATION IS BEING

ELONGATED AND ACTUALLY FROM THE

STANDPOINT OF USING PROFILING TO

LOOK AT THE EXPRESSION OF

PROTEINS, THIS IS A GREAT

SIMPLIFICATION BECAUSE IT MEANS

WE CAN LOOK AT THE OVERALL

DENSITY OF RIBOSE AND THE RATE

THEY ARE MOVING FORWARD IS VERY

SIMILAR ACROSS MESSAGES, WE CAN

GO DIRECTLY FROM THE RIBOSE FROM

COUNTING THE NUMBER OF FRAGMENTS

TO THE RATE OF PROTEIN SYNTHESIS

WHICH IS OF COURSE THE NUMBER

THAT IN GENERAL PEOPLE WOULD BE

MOST INTERESTED IN.

BUT IN DOING THIS EXPERIMENT, WE

HAVE GOT A GREAT GIFT AND THE

GIFT WAS THAT IN ADDITION TO

STOPPING INITIATION TRANSLATION,

THE HERRING TONE IN CAUSED

RIBOSE TO ACCUMULATE AT THE

POINT THE CODONS WHERE

TRANSLATION WAS STARTING.

SO YOU SEE THAT HERE, A BUILDUP

OF RIBOSE AS TREATED WITH

HERRING TONE IN AND THIS IS TRUE

GENERALLY AND ON GENE SPECIFIC

BASIS SO RIBOSOME FOOTPRINTS

ACROSS MESSAGES BUT THEN THEY

ACCUMULATED AT THE START SITE.

THE REASON WHY THIS IS SO

IMPORTANT IS THAT IF WE ARE

GOING IN A COMPLEX GENOME, IT

WOULD BE VERY HARD TO FIND ALL

THE PLACES BY A BIOINFORMATICS

APPROACH WHERE TRANSLATION IS

STARTING.

HOW DO YOU DISTINGUISH WHICH

ONES START AND WHICH ONES

AREN'T.

AS FROM DATA LIKE THIS ITS

UNAMBIGUOUS THIS IS THE DOMINANT

PLACE FOR TRANSLATION STARTING.

AND ONE CAN TAKE NO MESSAGES

WHERE YOU'RE CONFIDENT WHERE

TRANSLATION IS STARTING AND

PLACES WHERE YOU THINK IT ISN'T

AND USE THIS TYPE OF DATA TO

TRAIN A MACHINE LEARNING TOOL IN

OUR CASE A SUPPORT VECTOR

MACHINE WHICH IS A PRETTY PLAIN

APPROACH FOR DOING THIS, TO NOW

IDENTIFY NOVEL TRANSLATION START

TAKES AND WE DO THIS IN A GLOBAL

MANNER TO REALLY START NOW TO

TAKE THE HUMAN GENOMES WE HAVE

TO TURN THEM INTO DRAFTS OF THE

HUMAN PROTEOME.

WHEN ONE DOES THESE TYPE OF

MACHINE LEARNINGS, THE TYPICAL

APPROACH TO MAKE SURE YOU'RE NOT

FOOLING YOURSELF IS TO TAKE A

SUBSET OF THE DATA YOU'RE

LEARNING ON AND NOT LET THEM SEE

THAT AND HAVE THE SYSTEM OR

ALGORYTHM, AND THEN SEE HOW DO

YOU DO AT TREATING THIS?

YOU CAN SEE HERE IS AN EXAMPLE

OF THIS WHERE WE HAVE OR ARE

ABLE TO IDENTIFY THE ACTUAL

START SITES WITH ABOUT 85%

EFFICIENCY AND VERY NICELY

DISTINGUISH IT FROM THE

NEIGHBORING SITES WHERE

TRANSLATION IS NOT STARTING.

WE THINK THAT THIS 85% IS A 15%

OR OVER ESTIMATE OF HOW MUCH WE

ARE MISSING AND MORE LIKELY TO

REFLECT THE FACT THAT THE

ANNOTATION IS WRONG SOMEWHERE

CLOSE TO 15% OF THE TIME.

WE CAN NOW DO THIS IN

GENOME-WIDE WAY AND WE SEE THERE

IS MANY STARTS AS YOU EXPECT BUT

ALSO MANY PLACE WHERE IS

TRANSLATION WAS STARTING AT

PLACES WHERE THESE COGNATE SITES

THAT ARE OFF BY ONE NUCLEOTIDE

AND WHILE IT IS KNOWN THAT

TRANSLATION COULD START AT

NONCANONICAL STARTS, OUR ABILITY

TO ANNOTATE THIS IN A GLOBAL WAY

WAS VERY LIMITED.

WE ALSO SAW A DRAMATICzV

DIFFERENCE IN HOW THESE START

TYPES OF USED.

SO THESE NONCONKEL STARTS ARE

COMMON NOW UP STREAM REGIONS AND

WE THINK THAT MANY OF THESE ARE

LIKELY TO BE IMPORTANT IN THE

CASE HERE AND CRITICAL FOR

REGULATING THE TRANSLATION, RATE

OF PRODUCTION IN DOWNSTREAM.

WHEREAS MUCH MORE COMMON TO USE

AUG FOR CANONICAL OPEN READING.

EVEN AMONG THESE IN CODING

REGIONS, NEARLY HALF OF THESE

SITES WERE PREVIOUSLY

UNANNOTATED.

SO WE THINK NOW WE HAVE A NICE

SCHOOL FOR REALLY STARTING TO

LOOK BROADLY AT WHAT PROTEINS

ARE PRODUCED AND I'LL GIVE YOU

ONE OF MY FAVORITE EXAMPLES OF

THIS, THIS IS ACTUALLY FROM IN

THIS CASE FROM A VIRUS IN

FIBROBLASTS INFECTED WITH A

VIRUS AND YOU SEE THAT THERE IS

TWO ANNOTATED OPEN READING

FRAMES, ONE LONG ONE HERE BUT

YOU SEE NO FOOTPRINTS AND THAT

IS FINE.

IT'S PROBABLY JUST NOT EXPRESSED

UNDER THIS PARTICULAR CONDITION.

A SECOND ONE THAT IS WELL

TRANSLATED BUT THE REAL ACTION

IN TERMS OF TRANSLATION IS THIS

REGION IN BETWEEN THESE TWO

ANNOTATED OPEN READING FRAMES

AND YOU ASK WHY WASN'T THIS

ANNOTATED?

IT WASN'T FOR TWO REASONS.

FIRST IT DOESN'T START WITH THE

AEG.

IT STARTS WITH THE CUG WHICH WAS

QUITE SHORT.

IT'S ONLY BE25 AMINO ACIDS EVEN

THOUGH IT IS VERY NICELY

STARTING HERE AND SOFTENING HERE

AND CHASING WITH THE HERRING

TONE IN.

SO THEY START THE SHORT OPEN

READING FRAMES THAT'S BEEN

LARGELY MISFRIDAY ANY TYPE OF

ANNOTATION AND I'LL TELL YOU

HONESTLY, WE ONLY KNOW A VERY

SMALL HANDFUL OF CASES OF A

BIOLOGICAL ROLE OF THESE BUT I

CAN NOW WE HAVE AT LEAST THE

GROUNDWORK FOR STARTING TO

EXPLORE WHAT THEIR FUNCTION

MIGHT BE.

AND TO END THIS SECTION WITH A

BIT OF A ROGUE GALLERY OF THE

REANNOTATIONS.

A TRANSCRIPTION FACTOR THAT WAS

ANNOTATED STARTING AT AUG HERE.

IT STARTS IN UPSTREAM AND WHEN

WE WENT BACK WE SAW THAT IN THE

MAMMALIAN PROTEIN SOMEONE HAD

RUN A GEL ON THIS AND NOTED THE

PROTEIN SEEMED A LITTLE BIT

LARGE GIVEN ITS PROTECTIVE

SEQUENCE.

THERE IS A DRUNKATION WHERE THE

PROTEIN STARTS WHERE IT'S

SUPPOSED TO BUT A MAJOR FORM

CUTS OFF ABOUT A THIRD OF THE

PROTEIN AND THIS IS A

TRANSCRIPTION FACTOR AND THE

REGION THAT IS BEING CUT OFF

HERE IS A ACTIVATION DOMAIN SO

IT MIGHT TURN AN INHIBITOR INTO

AN ACTIVATOR OF TRANSCRIPTION.

AND HERE TWO PROTEINS, THIS

AGAIN IS FROM MOUSE EMBYRONIC

STEM CELLS NOW AND WE LOOKED AT

MYC, THEIR FAMOUS ROLE IN

PLURIPOTENCY AND YOU SEE THIS

VERY COMPLEX PATTERN OF

TRANSLATION BOTH IN UPSTREAM

OPEN READING FRAMES AND MULTIPLE

PLACES WHERE TRANSLATION CAN

START.

MYC IN PARTICULAR, THE EXTENT OF

THIS STUDY AS HAVING

TRANSLATIONAL CONTROL AND THE

CRITICAL ROLE OF THE REGION HERE

AND SO YOU CAN NOW SEE JUST HOW

COMPLICATED THE TRANSLATION IS

AND AGAIN, THIS IS JUST A

STARTING POINT FOR RELIEVING

THIS PHENOMENA TO THE BIOLOGICAL

CONTROL OF TRANSLATION IN THIS

CASE.

SO, REALLY I THINK THE BOTTOM

LINE FROM THIS PART OF THE TALK

IS NOW ANY HUMAN PROTEIN OR

MAMMALIAN PROTEIN YOU'RE WORKING

ON, YOU CAN START TO PULL UP

THESE SORTS OF FILES AND LOOK AT

YOUR INDIVIDUAL PROTEIN AND SAY,

WHAT IS GOING ON IN

TERMS -- WHAT IS BEING

SYNTHESIZED BY THIS GENE?

AND GIVE YOU A MINIMUM

GROUNDWORK FOR STARTING TO THINK

ABOUT ITS FUNCTION.

SO, I'LL NOW SWITCH GEARS AND

TALK ABOUT TWO VERY NEW

APPLICATIONS OF THE RIBOSOME

PROFILING APPROACH.

THE FIRST IS TO MONITOR CHAPERON

ENGAGED POLYPEPTIDES AND THE

SECOND GETTING AT A VERY CLASSIC

PROBLEM OF WHAT IS THE MEANING

OF SILENT MUTATIONS?

SO I'LL START WITH THE CHAPERON

STORY.

WHEN THEY MERGED FROM FOLDING

CHAINS HAS BEEN THAT THERE ARE A

VARIETY OF DIFFERENT TARGETING

FACTORS, PROCESSING FACTORS AND

CHAPERONS THAT ENGAGE THE CHAINS

AS SOON AS IT EMERGES FROM THE

RIBOSOME ACTIVE CHANNEL.

AND ALTHOUGH THE FACTORS OFTEN

CHANGE FROM PROCARYOTE TO

UCAREIOT, THE THEME YOU HAVE

THESE FACTORS THAT ARE DIRECTLY

RECRUITED TO THE RIBOSOME AND

CRITICAL FOR THE STATE OF THE

POLYPEPTIDE SEEMS TO BE A QUITE

PHENOMENON AND ALSO ANOTHER

THING THAT HAS COME OUT FROM

STRUCTURES OF THE RIBOSOME IS

THAT BIOCHEMICAL STUDIES, IS

THAT THIS EXIT PANEL, A VERY

CROWDED PIECE OF REAL ESTATE

WITH A VARIETY OF THESE

DIFFERENT TARGETING OR CHAPERON

FACTORS SEEMING TO BINDING TO

THE SAME SITE AND COMPETING WITH

EACH OTHER.

WE KNOW ALMOST NOTHING ABOUT THE

TRUE DYNAMICS IN VIVO SETTING

FOR HOW THE POLYPEPTIDE IS

RECOGNIZED AND YOU HOW THIS IS

ORG STRAIGHTED AND ON THE

RIBOSOME PROFILING APPROACH, WE

START TO GET AT THIS QUESTION.

AND FOR THESE STUDIES, WE

FOCUSED ON WHAT IS THE BEST

STUDY, THE BEST STUDY OF

RIBOSOME ASSOCIATE THE

CHAPERONS, THE TRIGGER FACTOR

HAS BEEN KNOWN FOR 20 YEARS.

IT'S BEEN STUDIED BOTH IN VIVO

AND BIOPHYSICALLY AND RECRUITING

BEAUTIFUL STRUCTURES OF THE

STRIGGER FACTOR BOUND TO

BACTERIAL OR PROKARYOTIC

RIBOSOMES SHOWING IT FORMS A

DOME OVER THE EXIT PANEL WAITING

FOR THE POLYPEPTIDE TO EMERGE.

SO OUR APPROACH FOR THIS IS A

SIMPLE MODIFICATION OF RIBOSOME

PROFILING.

INSTEAD OF SIMPLY LOOKING AT ALL

THE RIBOSOME PROTECTIVE FRAG

MINUTES IN ADDITION TO THAT, A

IMMUNOPRESIP TAILLIGHT THE

RIBOSOMES AND WITH THAT, THE

MASON CHAIN AND THE RIBOSOME AND

THE PROTECTIVE FRAGMENT AND

SEQUENCING THIS RIBOSOME

PROTECTIVE FRAGMENT, WE CAN SEE

PRECISELY WHEN AND WHERE THESE

DAISY CHAINS ARE BEING ENGAGED

BY THE CHAPERON.

SO A NICE ASSAY IN THE SENSE

THAT WE ARE UNDERLYING THIS

PROTEIN-PROTEIN INTERACTION.

HOW WOULD THE CHAPERON DEFINE

THE MASON CHAIN BUT WE HAVE A

READ OUT OF THE SEQUENCING WHICH

GIVES US THE PRECISION AND SPEED

ENABLED BY THAT TECHNOLOGY.

AT THE VERY FIRST LEVEL WE JUST

FOR EVERY MESSAGE, ASK HOW MANY

TRIGGER FACTOR MOLECULES WERE ON

THERE AND MADE A HISTOGRAM.

YOU CAN SEE THERE IS A VERY NICE

TRI-MODAL DISTRIBUTION FOR THE

DIFFERENT TYPES OF GENES.

SOME POORLY ENGAGED AND SOME

MUCH MORE STRONGLY ENGAGED.

WHEN WE LOOKED AT THE NAMES OF

THESE GENES, THEY ARE SPLITTING

THE PROTEOME BASED ON THE TYPE

OF PROTEINS SO INTERMEMBRANE

PROTEINS ARE LARGELY MUTANT FROM

RECOGNITION BY TRIGGER FACTOR.

SOLUBLE PROTEINS ARE ENGAGED IN

INTERMEDIATE LEVEL AND THEN THE

TIGHT BINDERS ARE THE OUTER

MEMBRANES PROTEINS AND WITH THIS

IT LED TO THAT THIS THE
TRIGGER

FACTOR SHAPER OWN MAY!XAVE

SOMETHING TO DO WITH THE

BIOGENESIS OF OUTER MEMBRANE

PROTEINS AND WITH THAT AND WHAT

DIRECTED EXPERIMENTS WE WERE

ABLE TO SHOW THAT IT IS THE

CASE.

AND I COME FROM A COACHING

CHAPERON OLD AND I GET

PARTICULARLY HAPPY ABOUT THIS

BECAUSE IT WAS AN EXAMPLE WHERE

PEOPLE HAVE BEEN STUDYING THIS

PROTEIN FOR 20 YEARS AND IT WAS

REALLY FAR FROM CLEAR WHAT ITS

MAJOR BIOLOGICAL ROLE WAS AND

HERE WAS AN OHMICS EXPERIMENT

THAT WHERE LOTS AND LOTS OF DATA

JUST WENT THROUGH A VERY SIMPLE

HYPOTHOSIS AND LED TO SOMETHING

THAT IS THE ROLE OF TRIGGER

FACTOR IN BIOGENESIS IN

IMPORTANT CLASS OF PROTEINS.

BEYOND JUST BROADLY SAYING

WHETHER TRIGGER FACTORS ON A

PARTICULAR MESSAGE OR OFF, THE

DATA LETS YOU SEE WHEN OR WHERE

IN THE SYNTHESIS OF THE

POLYPEPTIDES IS TRIGGER FACTOR

ENGAGING.

AND WHAT WE FOUND WAS THAT

REGARDLESS OF THE PROTEIN WELL

ENGAGED OR POORLY ENGAGED, THE

EXTREME TERMINIS FOR ABOUT

50-100 AMINO ACIDS WAS IMMUNE TO

RECOGNITION BY TRIGGER FACTOR.

AND WE COULD SHOW THAT THIS

WOULD NOT JUST A CORRELATION BUT

TRY TO GET A MORE DIRECT CAUSE

AND EFFECT BY TAKING ONE

SUBSTRATE IN THIS CASE, AND

TRUNCATING IT AT 5 PRIME END OR

EXTENDING OR MUTATING THE 5

PRIME END.

AND AS WE TRUNCATE IT, FURTHER

AND FURTHER FROM THE 5 PRIME

END, THE TRIGGER FACTOR COMES ON

LATER AND LATER.

WE ADD MORE ON, AND THE TRIGGER

FACTOR COMES EARLY.

IF WE CHANGE THE SEQUENCE, IT

DOESN'T MATTER.

IT DOESN'T EFFECT WHAT TRIGGER

OR WHEN TRIGGER FACTOR ENGAGES.

THAT'S SAYING THAT WHAT TRIGGER

FACTOR IS DETERMINING IS THE

LENGTH NOT THE ENGAGEMENT IS

REAL THEY THE LENGTH OF THE

POLYPEPTIDE SAYING THAT THIS IS

50-100 AMINO ACIDS IMMUNE.

AND THIS IS PLENTY OF

POLYPEPTIDES FOR THE MASON CHAIN

TO BE EXPOSED TO THE OUTSIDE

WORLD AND IT'S SATURDAYING THAT

THERE IS THIS PERIOD IN WHICH

THE CHAIN WOULD BE ABLE TO BE

PREFERENTIALLY ACCESS THE BY

OTHER FACTORS LIKE THE MODIFYING

ENZYMES SIGNAL RECOGNITION

PARTICLES THAT TARGET THE

PROTEIN AND THAT THERE MIGHT BE

REALLY IN VIVO COMPETITION AND

THAT AGAIN THE SUGGESTION IS

THIS COMPETITION WOULD BE

IMPORTANT TO MAKE SURE THE

TRIGGER FACTOR DOESN'T COME ON

TOO EARLY TO PREVENT THESE OTHER

CRITICAL STEPS.

THIS IS AN IDEA WE COULD TEST BY

OVER EXPRESSING TRIGGER FACTOR

AND SEEING HOW THIS EFFECTS

THESE MODIFICATIONS AND THE FACT

THAT A MODEST EXCESS IN VIVO OF

THIS CHAPERON CAUSED CELLS TO BE

VERY SENSITIVE TO INHIBITORS OF

THESE MAP MODIFICATIONS.

AGAIN, SORT OF PROVIDING IN VIVO

BACK IN THIS IDEA THAT THERE IS

A TRIAGING OF THE POLYPEPTIDES

AS SOON AS IT EMERGES AND A REAL

LOGIC TO HOW THE POLYPEPTIDES

ARE ENGAGED BY CHAPERONES OF THE

BEYOND THE SIGNIFICANTS OF THE

PARTICULAR RESULTS OUR

UNDERSTANDING OF THE BACTERIAL

TRIGGER FACTOR, I HOPE THAT THIS

GENERAL STRATEGY CAN BE USED FOR

BOTH UCARROTTICS AND PRO

CARROTTICS FOR UNDERSTANDING THE

LOGIC FOR DIFFERENT FACTORS FOR

WHEN THEY COME OFF.

SO, I'LL LEAVE THIS PART OF THE

TALK WITH ONE IMAGE.

THERE IS LOTS OF HUNGRY MOUTHS

WAITING TO FEED ON THE MASON

CHAIN BUT LIKE A GOOD PARENT,

THE RIBOSOME MAKES SURE THAT

EACH GETS FED AT THE RIGHT TIME

IN THE RIGHT WAY.

WITH THAT, I'D LIKE TO TALK

ABOUT THE FINA$M PART OF THE

MEANING OF SILENT MUTATIONS AND

I SHOULD SAY WORK OF A REALLY

TERRIFIC POSTDOCTORAL FELLOW WHO

HAS ONLY BEEN HERE FOR ABOUT A

YEAR AND MADE REMARKABLE

PROGRESS AND BUILDING ON STUDIES

FROM BACTERIAL RIBOSOME

PROFILING AND THE TRIGGER FACTOR

STUDIES.

AND THE BACKDROP TO THESE

STUDIES TO BE EXPERIMENT THAT I

HOPE WILL BE FAMILIAR TO MANY OF

YOU, A CLASSIC STUDY OR

EXPERIMENT IN MOLECULAR BIOLOGY

OF COURSE DONE NOT FAR FROM THIS

LECTURE, THE STUDIES OF

NEUREMBURG WHO 50 YEARS AGO

ALMOST TO THE DAY FIRST

PUBLISHED THAT IF ONE TOOK

POLYMERS AND PUT IT IN A

TRANSLATION EXTRACT, WHAT WOULD

BE TRANSLATED WOULD BE A POLYMER

OF ALANIN.

EXPERIMENTS WERE IMPORTANT IN A

NUMBER OF WAYS SO NOT THE LEAST

OF WHICH IS IT SAYS THAT THE

MAPPING FROM MRNA FOR THIS

MESSAGE TO POLYPEPTIDES WAS

UNIQUE AND THAT IS CRITICAL FOR

MESSAGES TO ENCODE INFORMATION

ABOUT PROTEINS.

YOU CAN'T HAVE A MESSAGE MAKING

MANY DIFFERENT PROTEINS AND HOPE

TO CONVEY INFORMATION.

BUT WHAT MIGHT NOT BE AS WELL

APPRECIATED BROADLY IS VERY

SHORTLY AFTER THIS EXPERIMENT,

THEY USED THE SAME BASIC

STRATEGY TO ASK THE QUESTION

ABOUT WHETHER THIS MAPPING IS

DEGENERATE OR NOT.

AND THE WAY THEY DID THIS WAS TO

REPEAT THE EXPERIMENT AND THIS

TIME ADD A LITTLE BIT OF C OR G

AND THEN ASK WHAT AMINO ACID

GETS INCORPORATED INTO THE

SYNTHESIZED POLYPEPTIDE?

AND THE MESSAGES WERE MOSTLY

POLY.

BUT IN ADDITION, WHETHER C OR G

WAS ADDED, THEY SAW THAT LEUCIN

WAS BEING INCORPORATED INTO THE

PROTEIN AND IT HAS TO BE ONE OF

THE MORE IMPORTANT RESULTS TO BE

PUT IN THE LEGENDS TABLE, THEY

NOTED THAT -- THEY DIDN'T HAVE

THE TERM CODON YET, BUT THE

CONCEPT.

THEY CONTAINED DIFFERENT

NUCLEOTIDES.

THE CODE IS PARTIALLY DEGENERATE

AND WE KNOW THAT THAT IS NOT

JUST TRUE IN THIS IN-VITRO

SYSTEM AS THEY CAUTIONED THERE

BUT IT'S BROADLY TRUE THERE ARE

61 DIFFERENT TRIPLET THAT IS

DON'T OR ENCODE THE PROTEINS BUT

20 AMINO ACIDS AND SO THAT EACH

AMINO ACID IS ENCODED BY

ANYWHERE BETWEEN ONE-6 DIFFERENT

CODONS AND IN THE CODE MEANS

THERE WAS A CHOICE AND ALSO

STARTED A 50-YEAR CONVERSATION

ABOUT WHAT THE MEANING OF THESE

CHOICES AND WHAT THE ORIGIN OF

THESE CHOICES WERE.

IT BECAME CLEAR EARLY AND

REMAINS CLEAR THERE IS

INFORMATION IN THE CODONS.

SO SOME ARE USED MORE COMMONLY

THAN OTHERS UNDER DIFFERENT

CONDITIONS.

NOT ONLY ARE INDIVIDUAL CODONS

USED MORE COMMONLY BUT SOME

PARIS OF CODONS, EVEN CORRECTING

FOR HOW OFTEN THEY ARE USED, ARE

USED MORE OR LESS OFTEN.

SO IT'S AN EXAMPLE WHY THEY ARE

ENCODED BY GGA OR GGU OR

GGU-GGA.

BUT GGA-GGU IS FAR LESS COMMON

THAN THE REVERSE.

WHY IS THAT AND WHAT IS THE

DRIVING FORCE?

ONE THING IS CLEAR THAT THERE IS

STRONG EVOLUTIONARY FORCE LIKELY

TO BE AN ANSWER.

AND ONE IDEA THAT IS OUT THERE

AND I'LL WARN YOU THIS IS A

STRONG AMENDMENT AND MANY PEOPLE

WHO ARE THINKING ABOUT THIS

WOULD AGREE, THAT NONETHELESS,

IT'S THE IDEA THAT HAS COMMON

CURRENCY, IS THAT I CALL IT

HYPOTHESES THAT LOW ABUNDANT

CRNA IS THE IN TRANSLATION.

FROM THE REASONS THAT THIS ISN'T

TOTALLY CRAZY IS THAT TRNAs

ARE VERY DRAMATIC IN HOW

ABUNDANT THEY ARE IN THE CELLS.

AND THEN IF YOU RECODE A MESSAGE

TO AVOID USING RARE TRNAs, IT

CAN INCREASE THE RATE OF

TRANSLATION.

SO THE NOTION THEN WOULD BE THAT

THE DRIVING FORCE BEHIND

MUTATIONS IS TO AVOID USING

THOSE SOLELY DECODED TRNAs

USING RIBOSOMAL RESOURCES.

THAT DOESN'T SOLVE THE QUESTION

BECAUSE THEN WHY ARE SOME

TRNAs RARE?

NEVERMIND WE ARE NOW IN POSITION

TO DIRECTLY ADDRESS THIS

QUESTION USING THE SAME RIBOSOME

PROFILING APPROACH.

IN THIS CASE, WE ARE GOING OFF

OF BACTERIAL DATA WHICH WILL

BECOME IMPORTANT AS YOU WILL SEE

IN A MINUTE.

ALTHOUGH THE SAME APPROACH CAN

BE USED IN THE MAMMALIAN DATA AS

WELL.

HERE WE SLASH THROUGH HAPPILY

GROWING E.COLI, MAKE RIBOSO

PROTECTIVE FRAGMENTS AND

SEQUENCE AND PLOT DENSITY AND

THE NOTION WOULD BE THAT

RIBOSOMES ARE PILING UP AND YOU

SEE THAT AS A PEEK IN HISTOGRAM

DENSITY OF RIBOSOME PROTECTIVE

FRAGMENTS.

THERE ARE LOTS OF POSITIVE

CONTROLS INCLUDING FROM THE

STOCK CODON END WHERE THERE IS

CRITICAL ROLE FOR RIBOSOMES IF

REGULATING TRANSLATION OF THE

DOWNSTREAM OPEN LEADING FRAME

AND IT'S CLEAR FROM THE DATA

WHERE THIS IS.

BUT YOU ALSO SEE OVER THE ENTIRE

MESSAGE THERE IS A HIGH

VARIABILITY IN RIBOSOME DENSITY

ACROSS MORE MESSAGE AND THIS

ISN'T TRUE JUST FOR THIS, IT'S

QUITE BROADLY FOR DIFFERENT

MESSAGES AND NOW RAISES OR LETS

US ADDRESS THIS QUESTION.

CAN THEY EXPLAIN THE CAUSES.

SO THIS FOR EXAMPLE A LOW

ABUNDANCE GNR OR A HIGH

ABUNDANCE GRNA THAT IS BEING

DECODED?

WHAT JEAN DID IS TOOK EUGENE'S

DATA AND TAUGHT IT TO THE

AVERAGE OCCUPANCY OF THE

RIBOSOME AT EACH OF THE 61

DIFFERENT CODONS AND AS A

FUNCTION OF THE NO ONE

ABUNDANCES OF THE TRNAs UNDER

THESE CONDITIONS.

IF YOU SQUINTED, YOU WOULD START

TO BELIEVE.

IT'S NOT A PERFECT CORRELATION

BY ANY MEANS BUT MAYBE THERE IS

RARE GRNAs THAT ARE BEING MORE

SLOWLY TRANSLATED.

THAT MUST GO UNTIL JEAN LOOKED

AT WHAT THESE WERE.

THEY ALL CODED FOR SEARINE AND

THEN HE READ UP ON THIS AND TOLD

ME SOMETHING I DIDN'T KNOW,

WHICH IS THE FIRST THING E.COLI

DO WHEN AT A RUN OUT OF GLUCOSE

IS THEY START USING SEARINE AS

THEIR CARBON SOURCE, NATURAL

SOURCE.

SO IT CLEARLY SEEMS LIKE IT WAS

LIKELY TO BE DUE TO TOO LITTLE

SEARINE.

WE WENT WITH MORE GLUEICOSE AND

IT WENT AWAY.

IT'S A NICE CONTROL.

WE CAN SEE THE RIBOSOME WHEN THE

TRNA IS DEPLETED BUT IT

COMPLETELY OBLITERATED THIS IDEA

WHAT WE ARE SEEING HAS ANYTHING

TO DO WITH THE TRNA ABUNDANCE.

SO THEN WHAT IS IT?

BASICALLY WHAT IN SUMMARIZING,

THE POSITION OF THE CODON AND

THE SITE WHERE THE IN COMING

TRNA WOULD BE COMING IN TO ADD

THE NEXT AMINO ACID, IS NOT

OVERALL AT LEAST CORRELATED WITH

HOW LONG֜ IT TAKES TO
TRANSLATE.

SO GENE CAME FROM A PHYSICS

BACKGROUND SO HE DID WHAT A GOOD

PHYSICIST WOULD DO, HE DID A

CROSS CORRELATION BETWEEN THE

SEQUENCE ACROSS THE RIBOSOME AND

PAUSING.

HE LOOKED TO THE T SITE, NOTHING

AT ALL.

HE LOOKED AT THE E SITE AND

NOTHING AT ALL AND THEN HE GOT

TO NUCLEOTIDES UPSTREAM IN THE

MESSAGE AND REALLY HIT PAY DIRT.

NOW FOR A VARIETY OF DIFFERENT

TRIPLETS, WE SEE A VERY STRONG

CORRELATION BETWEEN THE PAUSING

OF THE RIBOSOME HERE AND THE

IDENTITY OF THE TRIPLET HERE.

JEAN DID ONE STEP FURTHER, HE

STARTED TO PIECE THESE TRIPLETS

TOGETHER AND NOTICED THAT YOU

COULD FROM THEM STRING TOGETHER

A SHINED SEQUENCE AND TO REMIND

YOU, THOSE OF YOU WHO AREN'T UP

ON THE TRANSLATION, THE WAY YOU

START TRANSLATION ARE DETERMINED

IN E.COLI IS THROUGH AN

INTERACTION BETWEEN SO-CALLED

SHINE ORGAN OWER SD SEQUENCE IN

THE MESSAGE AND AN ANTI -- IN

THE RNA IN THE SMALL SUBUNIT.

SO THIS PARKS THE SMALL SUBUNIT

HERE AND IF THERE IS AN

INAPPROPRIATELY START SITE,

TRANSLATION CAN INITIATE FROM

THERE AND SO THIS THEN LED TO

THE HYPOTHESES IT WAS THE

PAIRING BETWEEN THE MRNA AND THE

RIBOSOMAL RNA THAT WAS

IMPORTANT.

AND NOW WHEN HE LOOKED AT THIS

PAUSING AS A FUNCTION TO THE

SO-CALLED AFD SEQUENCE, A VERY

STRONG ROBUST CORRELATION, ONE

SEIZE THIS IN THE NEGATIVE

BACTERIA IF E.COLI AND DISTANT

RERELATED GRAND PAUSEAL

BACTERIAL BUT IN THE UCAREIOT

THAT DOESN'T USE THIS MECHANISM

IN TRANSLATION INITIATION

DEMONSTRATING THIS ISN'T SOME

TERRIBLE ARTIFACT OF CLONING OR

SEQUENCING PROCEDURE BUT A

BROADLY CONSERVED MECHANISM FOR

TRANSLATION IN BACTERIA.

NOW, IF YOU STEP BACK, IMPLICIT

IN ALL OF THIS HAS BEEN IMPLIED

OR SUGGESTING WHAT IS GOING ON

THAT TRANSLATION OF ONGOING

TRANSLATION IS SLOWING DOWN WITH

THE POLYPEPTIDE HANGING OUT.

BUT AT LEAST AT THIS POINT,

CONSISTENT WITH THE DATA I HAVE

SHOWN YOU IS RATHER THAN

STALLING ONGOING TRANSLATION,

THE SECOND RIBOSEM IS COMING

HERE SEEING A SEQUENCE THAT

LOOKS LIKE A START SITE,

STARTING ASSEMBLING THE LARGE

SUBUNIT AND THEN BY SOME QUALITY

CONTROL, ANOTHER MECHANISM

REALIZING THIS ISN'T THE TRUE

PLACE TO START AND ABORTING

INITIATION.

AND SINCE WE ARE JUST LOOKING AT

RIBOSOME FOOTPRINTS, THESE COULD

GIVE YOU ACCESS OF RIBOSOME

FOOTPRINTS HERE.

SO, TO GET AT OR DIFFERENTIATE

THIS QUESTION, JEAN WANTED TO

UNDERSTAND, IS THE RIBOSOME

STALLING HERE THE SAME ONE THAT

STARTED HERE?

TO DO THIS HE USED A BIT OF

TECHNOLOGY THAT IS BEING

DEVELOPED BY A NUMBER OF LABS IN

MOST RECENTLY BY JASON CHIN

WHERE YOU CAN MAKE A SECOND COPY

OF THE RIBOSOMAL RNA THAT HAS

FAR MORE THAN A RIBOSOME THAT

OTHERWISE IDENTICAL THAT HAS A

DIFFERENT ASD SEQUENCING IN THIS

RNA.

NOW IF YOU EXPRESS THEM AS YOUR

ONLY RIBOSOME, CELLS WOULD BE

DEAD AS A DOORNAIL.

BUT IF YOU CO-EXPRESS IT ALONG

WITH THE WILDTYPE, REMARKABLY

ENOUGH THE CELLS DO QUITE WELL

AND ABLE TO SHOW THAT THE RILED

TYPE RIBOSOMES WILL START AT THE

NORMAL SEQUENCES AND RIBOSOMES

WILL ONLY START AT THE

ORTHOGONAL SEQUENCES.

SO NOW WE CAN DISTINGUISH

BETWEEN THESE MODELS BY
CHANGING

IN ONE PARTICULAR MESSAGE THE SD

OR THE SD START AND THE IDEA IS

THAT IF THIS MODEL WERE TRUE,

REGARDLESS OF WHICH RIBOSOME

STARTED, YOU STILL SEE INTERNAL

FOOTPRINTS BECAUSE THERE IS

PLENTY OF WILDTYPE RIBOSOMES

AROUND.

IF THIS MODEL IS TRUE, IT'S A

RIBOSOME PAUSINGS HERE TO BE

CHANGED SEQUENCE USED ORTHOGONAL

RIBOSOMES NO LONGER SEE THESE

SPIKES AND DEFINITIVELY ABLE TO

SHOW THIS TOP MODEL CAUSING OR

ONGOING TRANSLATION THAT WAS

BOTHERING.

SO IF YOU TAKE THIS, IT

COMPLETELY LOSES CORRELATION

WITH THE CD SEQUENCES AND GAVE

CORRELATION WITH THE SD STAR AND

IN THIS EXPERIMENT YOU HAVE 4000

INTERNAL CONTROLS OF ALL OF THE

ENDOGENOUS MESSAGES THAT STILL

PAUSE AT THE SAME SITE.

SO TO SUMMARIZE, WE HAVE NEW

IDEAS OF CODON PREFERENCE AT

LEAST SOME CODON PREFERENCE.

SOME CODONS INTERFERE WITH

TRANSLATION BY CAUSING -- AND

MESSAGES ARE PROBLEMATIC CODONS

AND THEN THE CELLS SAVES

RESOURCES BY NOT MAKING EXPRESS

OF THESE RARELY USED TRNAs.

SO YOU STILL GET THE CORRELATION

THAT IS TRUE IT'S WE HAVE THE

CAUSE AND EFFECT IN THIS SUBSET

THAT IS REVERSED.

IF THIS IS THE CASE WE SHOULD

SEE IMPACT, IF THIS IS TRUE, AN

IMPACT ON THE EVOLUTION OF CODON

CHOICE AND IN FACT WE DO.

WE LOOKED AT ALL THE HEX MERES

IN OPEN READING FRAMES THAT

THERE IS IN THE E.COLI, THERE IS

A STRONG DISENRICHMENT FOR ALL

THE HEX MERTHAT IS HAVE STRONG

AFFINITIES AND YOU CAN NOW SEE

HOW THIS PLAYS OUT NOT ONLY IN

CODON USAGE BUT IN PAIRS OF

CODON USAGE FOCUSING ON FLY SEEN

PAIRS.

SO THERE ARE 16 PAIRS AND YOU

SEE THE CURRENTS 16 DIFFERENT

WAYS THE CELL COULD CHOOSE TO

ENCODE AND A VERY STRONGLY

PREDICTED BY THE AFFINITY TO IN

PARTICULAR THE TENDENCY TO CAUSE

RIBOSOAPS TO PAUSE.

SO GETTING BACK TO THIS

QUESTION, I POSED OF WHY GGA,

GGT, IS FAR LESS COMMON THAN GGT

AND GGA.

WE CAN SEE THAT GGA AND GGT IS A

PERFECT MATCH TO THE SD TO THE

ANTISD.

SO JUST TO SUMMARIZE, THE

SD-HIKE FEATURES ARE MAJOR

DETERMINING ELONGIGATION RATE IN

BACTERIA AND AVOIDING THESE

FEATURES A DRIVING FORCE IN

CODON PAIR SELECTIONS.

IT DOESN'T EXPLAIN ALL THE

PHENOMENA BUT DOES PROVIDE A

CONCRETE EXPLANATION FOR

IMPORTANT SUBSET AND THEN

FINALLY IN ANALYSIS, I

DID -- DIDN'T HAVE TIME TO GO

THROUGH, THE PAUSES, GENERALLY

SELECTED AGAIN, WHAT THEY DO IS

THEY ARE OFTEN BROADLY CONSERVED

AND THAT SUGGESTS THAT THEY HAVE

A FUNCTIONAL ROLE.

AND SO, THIS NOW GIVES US A

FRAMEWORK IN THE VERY PRACTICAL

WAY TO START THINKING ABOUT WHEN

WE MAKE RECOMBINANT PROTEINS HOW

TO RECODE THEM.

SO, JUST A FINAL SLIDE AND A FEW

POSSIBLE FUTURE APPLICATIONS OF

THE RIBOSOME PROFILING APPROACH.

I WOULD ARGUE BASICALLY ANY

CURRENT USE FOR MRNA ABUNDANCE

MEASUREMENTS CAN BE COMPLIMENTED

BY OR REPLACED BY RIBOSOME

PROFILING AND LOSS OF ATTRACTIVE

EXAMPLES, LEARNING CANCER AGING

AND NEURODEGENERATION AND

MICRORNA TARGET DEGENERATION.

AND I THINK THIS WILL BE A

IMPORTANT FOR DEFINING HUMAN

PROTEOMES, IT'S NOT A FINAL STEP

BUT AT LEAST IMPORTANT STEP IN

HUMAN GENOME PROJECT.

WE CAN REITERATE THIS TO

DETERMINE WHICH PROTEINS ARE

MADE IN A GIVEN TISSUE OR IN

SUBSTUDY LIKE TRANSLATION AND WE

CAN DO SELECTIVE PROFILING AS I

HAVE SHOWN TO MONITOR PRO

TRANSLATIONAL ENGAGEMENT OF

CHAPERONS AND BROADLY WE HAVE A

TOOL NOW FOR LOOKING AT EXAMPLES

IN STARTING TO GET AT LEAST

CLUES IN THE MECHANISM OF HOW

TRANSLATIONAL CONTROLS IS USED

TO REGULATE EXPRESSION GENES AND

THEN WE HAVE NOW DATA TO START

UNDERSTANDING THE BIOLOGICAL

FUNCTION AND DETERMINING

RIBOSOME PAUSING.

SO WITH THAT, I'LL FINISH AND

JUST TAKE A MOMENT TO THANK THE

PEOPLE WHO REALLY DID ALL OF

THIS WORK.

THE REAL HERO OF THESE STUDIES

IS A FORMER POSTDOC, NICK, WHO

STARTED ALL THE RIBOSOME

PROFILING.

HE IS NOW AN INDEPENDENT LAB AT

CARNEGIE INSTITUTE.

HE IS A NEIGHBOR OF YOURS AND I

WOULD URGE YOU IF YOU'RE

INTERESTED IN THIS TO CONTACT

HIM.

ALTHOUGH HE PROBABLY WILL KILL

ME FOR SAYING THAT.

HE DID ALSO THE CELL STUDY.

GREAT CONVERSATIONS WITH THE

FOLKS AT LUMINA.

AND THE MYOSIS EXPERIMENTS AND

THE TRIGGER FACTORS WITH EUGENE

A STUDENT IN MY LAB AND BECKER,

A STUDENT ALSO WORKED

HAND-IN-HAND.

WE GOT A LOT OF ADVICE AND

REAGENTS AND HELP FROM CAROL

GROWS'S LAB.

SO WITH THAT, I'LL THANK YOU

ALL.

FINISH AND TAKE ANY QUESTIONS.

[APPLAUSE]

>> THAT WAS FANTASTIC.

WE HAVE MICROPHONES IN THE

AISLES.

IF YOU HAVE A QUESTION, PLEASE

USE THE MIC SO PEOPLE ON THE WEB

CAN HEAR.

WE'LL TAKE A FEW QUESTIONS.

I GUESS I WILL ASK THE OBVIOUS

QUESTION FOLLOWING UP THE LAST

PART OF YOUR PRESENTATION, OKAY,

YEAST AND OTHER EUKARYOTES DON'T

CARE ABOUT SHINE DEL BEGANO

BOXES.

WHAT DO YOU SEE THERE THAT MIGHT

EXPLAIN PAUSING OF THE

RIBOSOMES?

IS THERE A TRNA ABUNDANT STORY

THERE EVEN THOUGH THEY DIDN'T

TURN OUT TO BE MUCH OF ONE IN

BACTERIA?

>> TRNAs ARE NOT MAJOR

DETERMINANT IN THE MAMMALIAN

DATA.

THERE IS SOME THAT WE SEE BUT IT

DOES A VERY CLEAR CANONICAL

SEQUENCE OF CAUSING PAUSING

GLUTAMATE IN MAMMALIAN CELLS BUT

IT SEEMS TO BE BROADLY USED.

BUT THERE ARE ALSO A LOT OF

SPECIFIC EXAMPLES QUITE ROBUST

PAUSING THAT YOU'LL SEE.

ABOUT ONE IN EVERY 4 MESSAGES IN

THE DATA HAS A ROBUST PAUSE,

25-50 FOLD SLOWER THAN THE

AVERAGE, AND JUST SCRATCHING THE

SURFACE ON THAT BIOLOGICAL ROLES

ARE.

>> RELATED TO THE INITIATION

SIGNALS SEEN WITHIN CODING

REGIONS, DO YOU SEE ANY EVIDENCE

OF THE PROTEIN WITHIN A PROTEIN

POSSIBLY EVEN OFFSET?

OR ANY OF THESE SIGNATURES OF A

STAR TRANSLATION?

>> YES, SO WE HAVE THIS JUST

CAME OUT BUT IN THE
MAMMALIAN

DATA, WE HAVE MANY MESSAGE THAT

IS HAVE MORE THAN ONE START SITE

AND A MAJOR ONE AND MINOR ONE

OFTEN EVEN THE MAJOR ONE ISN'T

THE ANNOTATED ONE.

IN GENERAL, MOST OF THE TIME

ANNOTATION IS RIGHT.

>> BUT ARE THEY EVER OFFSET?

>> THEY ARE DEFINITELY SEE,

ESPECIALLY IN VIRAL WORK WE DID

ON A MEG LOW VIRUS.

YOU SEE THE MOST LARGEST HUMAN

VIRUS IS 250KB.

IT'S ABOUT HALF THE SIZE OF THE

SMALLEST BACTERIA AND IN THERE

YOU SEE TRANSLATION AND OPPOSITE

STRANDS, TRANSLATION IN MORE

THAN ONE READING FRAME AND ALL

NONCANONICAL IMPACTING.

IN THE MAMMALIAN WE SEE AN

OCCASIONAL ONE WHERE YOU SEE A

DIFFERENT FRAME BUT BY-AND-LARGE

IT'S TRUNCATIONS ARE ELONGATION.

>> JONATHAN, YOU DIDN'T SAY MUCH

ABOUT THE WORLD OF RNA STRUCTURE

INVOLVED IN TRANSLATION.

THERE ARE NICE EXAMPLES,

ESPECIALLY WITH RETROVIRAL

RNAs AND POSSIBLY DNAX THAT

COULD BE USED TO FIRST OF ALL

SHOW CONCLUSIVELY YOU CAN SEE

FRAMES SHIFTING WITH YOUR

METHODS AND SECONDLY LOOK AT THE

ROLE OF OTHER RNA STRUCTURES

PREDICTED TO CAUSE -- CONDUCIVE

TO FRAMESHIFTING.

>> OF COURSE TYLER YEARS AGO

TOOK THESE EXAMPLES IN THE

FRAMESHIFTING AND WHEN WE LOOKED

AT THE BIOLOGICAL EXAMPLE --

>>  [INAUDIBLE] --

>> VERY HIGH --

>>  [INAUDIBLE]

>> YES, SO YOU CAN

CLEARLY -- WHEN WE LOOKED AT

KNOWN EXAMPLES, AS FARES -- AS I

KNOW, WE HAD NO PROBLEMS SEEING

THEM EVEN WHEN THEY ARE SUB

STRIKING METRIC.

THERE IS -- THE KING OF THIS

DECODING HAS BEEN LAST WEEK

THREE DAYS IN THE LAB DOING SOME

PROFILING EXPERIMENTS TO GET

BROADER DATA ON THIS AND WE ARE

ALSO -- WE AND OTHERS HAVE

APPROACHES FOR GLOBALLY LOOKING 

IN VIVO AT RNA STRUCTURE USING

FOOTPRINTING TYPE OF STRATEGIES

NOW WITH THE DNA SEQUENCING READ

OUTS THAT WILL GIVE US EMPIRICAL

DATA ON WHERE THE STRUCTURE IS

TO GET AT THIS QUESTION.

>> AND THERE HAVE BEEN EFFORTS

TO FIGURE OUT WHETHER

FRAMESHIFTING IS A COMMON EVENT

IN TRANSLATING REGULAR GENES NOT

JUST VIRAL GENES AND YOUR

METHOD -- IS THE POSSIBILITY OF

LOOKING MORE CLOSELY AT THE

GENOME FOR FRAMESHIFTING --

>> ABSOLUTELY.

AND I WOULD SAY IN GENERAL, AS

WE GO TO STRESS CONDITIONS OR

STARVATION OR NONCANONICAL

CONDITIONS, EVERY PLACE WE

LOOKED IT BECOMES FAR MORE

COMMON.

SO IT IS A SENSITIVE AT THIS

POINT AND ANNOTATION PROBLEM, AN

ANNOTATION SOUNDS LIKE IT'S

MINIMIZING.

IT'S A BIOINFORMATICS PROBLEM

BECAUSE WE CAN COLLECT THE DATA

QUITE READILY.

AND VERSUS ALL OF THESE

EXPERIMENTS THERE IS AT LEAST

THREE QUARTERS OF THE TIME ON

THE COMPUTER AND NOT COLLECTING

THEM.

SO, ALL OF THIS IS STUDENTS OUT

THERE IS ADVERTISEMENTS TO TAKE

THE CLASS AND LEARN A BIT OF

BIOINFORMATICS BECAUSE IT IS

GOING TO BE INFORMATION OF

SCIENCE AND THAT WILL BE THE

WAY.

>> TO BUILD ON THAT, THE IDEA OF

LOOKING AND USING ALGARYTHMS FOR

PREDICTING OTHER KINDS OF FOLDS

MAYBE JUST AS IMPORTANT AS

LOOKING AT TRNA ABUNDANCE AND

HOMOLOGY.

>> RIGHT,ITERATING BACK AND

FORTH.

>> ANY INFORMATION ABOUT

MITOCHONDRIA?

THAT'S SOMETHING I'M INTERESTED

IN THIS STUDY.

ABOUT THE WHAT?

>> MET QUANDARYIA.

>> WE DO NOT HAVE ANY DATA ON

MITOCHONDRIA.

WE HALF HEARTEDLY TRIED TO GET

THOSE DATA.

I BELIEVE THAT THE AS AS SHOULD

BE READILY ADAPTABLE TO LOOK AT

MITOCHONDRIAL TRANSLATION BUT

YOU HAVE TO ADAPT THE ASSAY

SPECIFICALLY TO GET THE

FOOTPRINTS OF THE MITOCHONDRIAL

RNA AND INTO INHIBIT

MITOCHONDRIAL TRANSLATION

RAPIDLY.

SO IT'S A REALLY NEAT QUESTION.

IF ANYONE WOULD BE INTERESTED IN

DOING IT, I'M HAPPY TO

EVENTUALLY SELL IT TO SOMEONE IN

THE LAB, THOUGH.

>> WE HAVE LEARNED A LOT ABOUT

THE INITIATION SO JUST WONDER

FIGURE HAVE YOU SEEN ANY BIAS

FOR STOP CODONS?

>> BIAS TOWARDS STOP CODONS?

IN WHAT SENSE?

>> ANY OF THE CODONS WHICH

RECOGNIZES STOP CODONS?

>> CODONS -- SO NON- -- NOT

CANONICAL STOP CODONS?

>> YES.

>> THAT'S A HARDER QUESTION.

NOT QUANTITATIVELY.

IT'S VERY RARE TO SEE

QUANTITATIVE TERMINATION AT A

NONSTOP CODON.

IT'S A LITTLE -- IT'S HARD FOR

US TO SEE IF A 50% OR 25% OF THE

RIBOSOME STOPPED IN A PARTICULAR

SPOT BECAUSE WE DON'T HAVE THE

EQUIVALENT -- IF WE DID, WE

COULD FREEZE AND ONLY LOOK AT

THE RIBOSOMES STOPPING AND

ADDRESS THAT.

IN GENERAL, I SHOULD SAY THAT

THE DENSITY AFTER STOP CODONS IS

QUITE ROBUST.

SO BY-AND-LARGE, MOST RIBOSOMES

ARE STOPPING UNTIL YOU START TO

GET MORE CONDITIONS.

>> AGAIN, THERE WILL BE A

RECEPTION WITH COFFEE AND

COOKIES AND SUCH IN THE LIBRARY.

PLEASE JOIN THE SPEAKER THERE.

LET'S THANK JONATHAN ONE MORE

TIME.

[APPLAUSE]