A"semi'automatic"web"based"tool - - PDF document

! !

A"semi'automatic"web"based"tool for"the"selection"of"research projects"reviewers

Valeria!PUPELLA!a,!Maria!Eugenia!MONTEVERDE!a,!Claudio!LOMBARDO!b, Filippo!BELARDELLI!c!and!Mauro!GIACOMINI!a

a""""Department"of"Informa-cs,"Bioengineering,"Robo-cs"and"System"Engineering,"University"of"Genoa,"Via"Opera """"Pia"13,"16145"Genoa,"Italy b""""Alleanza"Contro"il"Cancro,"Via"Giorgio"RiboIa,"5,"00144,"Rome,"Italy c""""Is-tuto"Superiore"di"Sanità,"Viale"Regina"Elena"29,"00161,"Rome,"Italy

! !

Introduction

• New%postdoctoral%scien1sts%are%poorly%integrated%into%the%world
• f%research
• How% to% promote% their% research% projects% at% an% interna1onal

level?

• TRAIN%(Training%through%Research%Applica1on%Italian%iNi1a1ve),

a%new%mobility%program%was%conceived

• But$how$iden,fy$the$appropriate$reviewers?

And now?

A"web"tool

! !

A"publication"based"tool

A"reviewer"must"be"an"expert"of"his"chosen"field:"how"to measure"his/her"exper:se? "

Publica(ons

Informa:ons"about"relevance"and"impact"of"reviewer’s"research can"be"retrieved"from"his/her"publica:ons. "

! !

Data$collection

• TRAIN'foresees'three'different'mobility'schemes'for'scien8sts,

either' outgoing' or' ingoing' and' the' re=integra8on' into' the

• ncology'field.

'

TRAIN

Personal'Data 20'different research'fields

List'of'20'most relevant publica8ons'in'the chosen'fields (PubMed'ID)

The!reviewer!will!go!through!the!same!registra1on steps.

! !

Registration

! !

Insert'publications

!

! !

Research()ields

Keyword Descriptor.UI MeSh.Term Cu$ng!edge!and!innova-ve research!related!to transla-onal!oncology D057170 Transla-onal!Medical!Research Applica-ons!to!transla-onal

• ncology!coming!from!other

fields D057170 Transla-onal!Medical!Research Medical!Oncology D008495 Medical!Oncology Epidemiology!and!Cancer registries D004823 Epidemiology Imaging D003952 Diagnos-c!Imaging Radiotherapy D011878 Radiotherapy Biotherapy D001691 Biological!Therapy Pallia-ve!Care!and Rehabilita-on D010166 Pallia-ve!Care Pathobiology!(including Biomarkers) D015415 Biological!Markers

! !

Research()ields((2)

Keyword Descriptor.UI MeSh.Term Oncogenomics!and!Gene.cs D023281 Genomics Cancer!Immunology D000486 Immunology Mutagenesis!and Carcinogenesis D016296 Mutagenesis Bioinforma.cs!and!Structural Biology D019295 Computa.onal!Biology Pharmacogenomics D010597 Pharmacogenomics Molecular!and!Cell!Biology D008967!– D003585 ! Molecular!BiologycI!Cell!Biology MicroIinvasion!and Metastasis D009362 Neoplasm!Metastasis

! !

Why$PubMed$ID?

PubMed!ID

WoK MeSH!descriptor, Qualifiers Number!of!cita=ons PubMed E@U=lity Relevancy!with!the!research!field Impact!of!each!publica=on

! !

Semi%automatic+tool

• Automa'c:#data#retrieved#from#PubMed#Web#Service;

# # # # #

• Manual:#data#retrieved#from#Web#of#Knowledge.

! !

The$platform

Web%based!system:

• Most!feasible!solu4on!for!interac4ng!with!the!database;
• Flexible!and!low%cost!tool;
• Allows!independent!and!simultaneous!work!in!different

reali4es!all!over!the!world;

• Allows!immediate!data!sharing.

!

www.trainoncology.eu

! !

The$platform$online

! !

How$it$works

PubMed!ID MeSh!Descriptor Keyword!(research field) MeSh!Descriptor Relevance!control!tool Reviewer!is!expert!in the!field!OR!not AutomaBc

! !

How$it$works

Manual Publica+on Number!of!cita%ons Cita+on!related!index TRAIN!DB Publica+on!ranking

! !

Citation'related'index

Is#h$index#inspired. # It#represents#the#number#of#papers#(n)#which#have#been#cited#at least#a#number#(n)#of#;mes. # For#reviewers,#it#ranges#from#0#to#5,#since#they#can#upload#up#to 5#publica;ons.

! !

Selection)of)reviewers

Manual!step ! !

• Administrators!can!choose!the!best!reviewer!from!the

classifica7on!by!the!cita7on!related!index.

! !

Results

• 435"experts"took"part
• 45"did"not"upload"any"publica4ons,"as"a"demonstra4on"of

their"non9willingness"to"par4cipate"to"the"project.

• 9"research"areas"had"more"than"20"experts"registered,"the

match"algorithm"between"the"keywords"and"the"MeSH Descriptors"was"carried"out"in"order"to"level"the"different categories."For"the"other"fields,"it"was"not"necessary.

• Thanks"to"this"selec4on,"162"individuals"were"discarded.
• 23"out"of"the"remaining"228"did"not"have"a"sufficient"cita4on

related"index."The"established"threshold"was,"in"fact,"2.

• So,"accepted"reviewers"were"205.

"

! !

Conclusions

• The%evalua*on%tool%was%developed%in%order%to%obtain%a%higher
• bjec*veness%in%proposal%reviews.
• The%choice%of%reviewers%was%based%on%their%publica*ons%and

field%of%exper*se%and%not%only%on%their%candidacy.

• The%website%administrators%could%use%such%a%ranking%as%a

sugges*on%and%could%also%choose%autonomously%the%reviewers for%each%proposal.

• During%the%TRAIN%project,%340%reviews%were%carried%out;

therefore,%each%expert%evaluated%at%least%one%proposal.

• All%types%of%users%considered%the%tool%as%both%valuable%and

efficient.

! !

THANKS'FOR'YOUR ATTENTION!