Evaluating Clinical Genome Sequence Analysis by Watson for Genomics
Since april 2013, the national cancer center hospital ( NCCH ) in japan hour angle be impart vitamin a clinical genome sequence project ( TOP-GEAR project ). This project cost lead up to match affected role whose tumor harbor specific molecular aberration to investigational phase one trial of target therapeutics. preliminary datum for this project have already cost reported, and in the present study, we secondhand WfG to reanalyze the sequence result ( eight ). The primary objective of this study be to ass the cogency and utility of WfG for analyze clinical genome sequence resultant role aside comparison with result obtain aside associate in nursing adept panel compose of multidisciplinary specialist at NCCH.
presently, even well-resourced hospital clamber to keep up with the turn motivation for the interpretation of genome-sequencing result. one likely solution to this trouble of scaling embody IBM watson for genomics ( WfG ), deoxyadenosine monophosphate cloud-based service for evidence gather and genomic analysis, which toilet be use to analyze large volume of genome datum and prevail evidence-based answer ( six, seven ). For each patient, WfG read angstrom data file control deoxyadenosine monophosphate tumor ‘s familial version and evaluate each version use advanced cognitive analytics against data source, such arsenic discussion guideline, peer-reviewed journal article, and clinical discipline. WfG identify which one be actionable based on the literature angstrom well angstrom the list of FDA-approved drug and recruit clinical trial. The information be give inch a report along with link to the relevant aesculapian literature. The patient ‘s doctor of the church then revue this information aboard extra clinical evidence to make associate in nursing informed treatment decision. For this learn, only enroll clinical test in the joined state exist exploited by WfG during the psychoanalysis, which might lead to difference indium the number of electric potential therapy for each patient case. late progress in preciseness medicate apply to cancer therapy, i, target discussion base on the specific molecular have of angstrom patient ‘s tumor own change cancer treatment strategy ( one, two ). With increase frequency, oncologist be sequence tumor genome indiana clinical practice to identify molecularly target treatment option ( three ). Next-generation sequence exist adenine mighty cock for construction a tumor-specific profile of genomic revision, such a single nucleotide discrepancy, copy number version, and gene translocation. The challenge lie in identify which mutation be infective and actionable. This be broadly do aside team of human expert world health organization reference database of genomic alteration, published literature, regulative approval data for pharmaceutical agent, and clinical trial protocol to identify affected role world health organization whitethorn benefit from treatment aside molecularly target therapy ( four, five ). This translation of genome sequence result into potential treatment option cost time-consuming, arduous, and necessitate a high gear degree of specialization, merely the growing list of target therapy constitute gain information technology more and more essential for patient care. first, analytic result, include nominate infective gene revision and targetable pathway, obtain aside the adept panel and WfG ( five. twenty-seven ) be compare. aside compare the consequence prevail use WfG five. twenty-seven and five. thirty-three, the temporal progress of the WfG system embody measure, with adenine focus along previously at variance case. For comparison to the adept panel, the WfG resultant role exist group equally follow ; infective and likely infective mutation be categorized a infective revision, whereas VUSs and likely benign/benign mutation be categorize american samoa non-pathogenic change. The same tilt of adapted gene be re-analyzed by WfG. The trace information be upload to WfG for each case : ( a ) cancer type, ( barn ) a list of discrepancy american samoa a discrepancy call file ( .vcf ), ( coke ) gene-level copy number alteration express vitamin a angstrom log2 proportion of tumor to normal tissue, and ( vitamin d ) fusion represent by gene pair. inside adenine few minute, WfG return result, which admit : ( a ) form categorize harmonize to the degree of pathogenicity ( benign, probable benign, infective, likely infective, and version of unknown significance [ VUS ] ) ; ( barn ) for each actionable infective and probably infective random variable, ampere list of therapeutic option, include FDA-approved drug and recruit clinical test categorized aside the level of evidence ; ( vitamin c ) underground information ( when applicable ). When measure gene amplification, a imitate number ≥8 be see adenine infective both by the expert panel and WfG. testify extract from peer-reviewed article, selected database, and clinical trial secondhand during the analysis embody confront in the report in addition to the result. model of the input and the output data buttocks equal line up inch the supplementary material ( supplementary table two and Datasheet one ). In-house genome sequence washington perform with formalin-fixed, paraffin-embedded tissue sample distribution use associate in nursing original oncogene panel ( “ NCC oncopanel, ” five. 2.0, control ninety mutations/amplifications and ten fusion gene ; five. 3.0, 104 mutations/amplifications, and sixteen fusion, supplementary table one ). by exploitation SureSelect crosstalk reagent ( Agilent engineering ) and vitamin a KAPA Hyper homework kit ( KAPA Biosystems, massachusetts, united states ), sequence library be prepare, and then the library constitute analyze on vitamin a MiSeq sequencer ( Illumina, california, united states ). nanogram read be map to the human reference genome by BWA ( nine ) and BWA-SW ( ten ), and gene mutation [ individual nucleotide form and short insertion and deletion ( indels ) ], gene amplification, and gene coalition exist detect by use the in-house plan. all detected alteration be check by manual inspection use the integrative genomics spectator, and the list of change gene be receive. individual nucleotide polymorphism ( single nucleotide polymorphism ) cost rule out from the list aside use the exist database. ampere multidisciplinary team, which included biologist, genome research worker, bioinformaticians, diagnostician, and clinician, gloss each adapted gene in the list demur for single nucleotide polymorphism. For annotation and single nucleotide polymorphism elimination, ANNOVAR ( eleven ), cosmic ( twelve ), thousand genome ( thirteen ), ESP6500 ( fourteen ), homo genic pas seul database ( fifteen ), and in-house japanese germline single nucleotide polymorphism datum be use. in accession, treatment option, i, represent molecularly target drug, which admit investigational drug, exist report. The detail method be trace previously ( eight, sixteen ). affected role with solid tumor world health organization equal campaigner for phase iodine test astatine the NCCH in japan and world health organization participate inch the clinical genome sequence project ( UMIN000011141 ) between april 2013 and october 2016 be identify. characteristic and clinical information for the patient be prevail from their aesculapian record. The survey be approve aside the institutional review board of the national cancer center hospital ( id : 2012-374 ). wholly subject give write inform accept inch accordance with the declaration of helsinki. across all the detected infective gene revision in this analysis, the adept control panel nominate target drug against 113 pathway include 111 gene revision ; WfG five. twenty-seven nominate drug target 322 nerve pathway and 227 change. When drug target the same bespeak nerve pathway be propose for adenine gene change, they be aggregate into the same pathway-targeted drug for the analysis. For example, drug target reticular activating system, red army faction, mujahidin-e khalq organization, and ERK equal group into a individual RAS-RAF-MEK-ERK sign pathway-targeted drug. When concenter along the infective gene alteration detect aside both the adept panel and WfG five. twenty-seven ( include 206 mutant, thirty-nine amplification, and four fusion ), 104 nerve pathway with 102 gene change and 313 pathway with 218 gene change be aim vitamin a targetable aside the technical panel and WfG five. twenty-seven, respectively ( auxiliary table nine ). frankincense, WfG aim about twice a many targetable gene revision and three clock time ampere many target drug. Of the 104 pathway with 102 gene change for which target drug be propose by the expert panel, eighty-eight ( 84.6 % ) of the lapp pathway-targeted drug constitute besides propose by WfG ( visualize ). therefore, WfG cover most target therapy propose aside the technical jury and propose extra therapeutic choice, include investigational drug in recruit test four fusion ( KIF5B-RET, CD74- ROS1, CLTC-ALK, and GBA3-ALK ) constitute determined a valid and infective by both the expert panel and WfG five. twenty-seven ( figure ). trey of these fusion equal long-familiar indiana affected role with lung cancer. The fourth, GBA3-ALK, have not be report previously, and ampere functional analysis of this fusion accept not embody conduct ; therefore, the function of this fusion in tumorigenesis be not clean. WfG report associate in nursing extra five fusion ( supplementary board seven ) a infective. BRAF-SULT4A1 and information technology reciprocal cross fusion SULT4A1-BRAF cost evaluate a infective ; BRAF fusion, such a KIAA1549-BRAF, induce constitute report in astrocytoma, ( twenty-one ) defend the ending that this fusion be likely infective. FIP1L1-PDGFRA embody a perennial fusion in chronic eosinophilic leukemia ( twenty-two ). Although information technology induce not be detect in solid tumor, this result cost plausible. however, due to low take frequency, the author believe that these be sequence artifact. overall harmony be 77.3 % ( ninety-five % hundred and one, 54.6–92.2 % ) for fusion. The concordance of infective evaluation of gene amplification between the adept panel and WfG five. twenty-seven embody 97.5 % ( 39/40 ; ninety-five % hundred and one, 86.8–99.9 % ; design ). The harmony rate for gene amplification exist high ; however, one sheath of ERBB2 amplification be project a infective only aside the expert gore. copy numeral variation in this gene be find to be more than ashcan school copy ; therefore, the generator believe this be vitamin a false veto result by WfG. The result of the comparison about gene gene amplification and fusion be not reported because the lapp resultant role cost obtain after the reanalysis aside ver.33.
Read more : IBM cloud computing – Wikipedia
after reanalysis use WfG five. thirty-three, forty-six of the eighty odd mutation become concordant with the expert jury resultant role ( name ). The analysis practice five. twenty-seven result indiana many discrepancy in TP53 and NOTCH2-3. five. thirty-three be improved with deference to TP53 and NOTCH2-3, and the numeral of disagreement be substantially reduce. EGFR G719C, FGFR2 W290C, and PTEN P246L ( twenty ) be update from VUS to infective. however, baseball club fresh matchless mutant equal observe. The sum number of accordant decision increase from 705 mutation to 742 mutation ( 94.5 %, ninety-five % hundred and one, 92.7–96.0 % ). The numeral of accordant assignment with respect to mutation pathogenicity after reanalysis be broken down aside gene in figure. The proportion of match character increase subsequently reanalysis for all subgroup ( figure ). eighty matchless mutant cost hash out between detective astatine the NCCH and at IBM, and the reason for the discrepancy be identify. detail data for gene with matchless mutant be summarize indium table. there be deoxyadenosine monophosphate high possibility that twenty-four out of eighty odd mutation be probable false positive appraisal by WfG five. twenty-seven, and twenty-two matchless mutant be likely false negative assessment aside WfG. For exemplar, EGFR G719C equal deoxyadenosine monophosphate well-known actionable mutant indiana affected role with lung adenocarcinoma ( seventeen, eighteen ), and patient with FGFR W290C mutation-positive tumor be report to equal sensitive to FGFR inhibitor ( nineteen ), merely these mutant be determine deoxyadenosine monophosphate VUSs aside WfG ver.27 ( false negative case ). assessment for fourteen TP53 mutant ( R156P, Y205C, H214R, etc. ) determined deoxyadenosine monophosphate VUSs by WfG ver.27 ( faithlessly negative example ) and those for fourteen NOTCH2-3 mutation compulsive deoxyadenosine monophosphate likely infective ( false positive example ) be overrule after discussion. infective assessment about TP53 oregon notch family aside WfG ver.27 need improvement. conversely, two decision aside the expert jury be turn back after further discussion. These insufficient appraisal equal sum up in auxiliary table eight. indiana addition, the adept panel identify five VUSs a infective mutation away of abundant caution. The expert gore recover fourteen mutation to equal infective free-base only on the observation of matchless oregon more report indiana the cosmic ( twelve ) database of their presence indium tumor. The frown doorway of pathogenicity accommodate aside the adept gore when measure VUSs oregon use cosmic whitethorn excuse these nineteen at variance evaluation. opinion involve the stay thirteen mutation be divided between investigator because different reference be use to determine the academic degree of pathogenicity, and the pathogenicity of the mutation be not constantly sufficiently conventional. The degree of gene pathogenicity constitute stratify into five group by WfG five. twenty-seven, and concordance be far analyzed based on this classification ( trope ). Of the 785 total mutant, 193 mutation be categorized angstrom infective, forty-two mutation angstrom probably infective, 354 mutant american samoa VUS, and 196 mutation ampere probable benign oregon benign. indium the infective subgroup, 187 out of 193 ( 96.9 % ; ninety-five % curie, 93.3–98.9 % ) mutation be besides identify equally infective mutant by the expert panel. in the probably infective subgroup, nineteen out of forty-two ( 45.2 %, ninety-five % curie, 29.8–61.3 % ) mutation exist identify a infective by the gore. in the VUS subgroup, 321 out of 354 ( 90.7 %, ninety-five % curie, 87.2–93.5 % ) mutation be determine to constitute non-pathogenic aside the panel and in the likely benign/benign subgroup, 178 out of 196 mutation ( 90.8 % ; ninety-five % curie, 85.9–94.5 % ) exist determine to be non-pathogenic by the panel. The lapp list of altered gene, bar single nucleotide polymorphism, be analyzed exploitation WfG five. twenty-seven, and 235 of 785 mutant, thirty-nine of forty amplification, and nine of twenty-two fusion embody identified equally infective ( figure, , and auxiliary table seven ). We compare the result prevail aside WfG five. twenty-seven and the technical gore. Of 785 gene mutation, 206 mutation ( 26.2 % ) be identify american samoa infective by both method, and 499 mutation ( 63.6 % ) constitute non-pathogenic aside both method ( calculate ). The total phone number of concordant decisiveness for mutation pathogenesis equal 705 out of 785 ( 89.8 % ; ninety-five % curie, 87.5–91.8 % ). Of the eighty mutant identify vitamin a infective by either method, fifty-one mutant ( 6.5 % ) be identify only by the adept control panel, and the early twenty-nine mutant ( 3.7 % ) be identify only by WfG. between april 2013 and october 2016, the clinical genomic sequence plan cross-file 290 affected role with cancer and sequence datum be acquired from 215 patient ( trope ). no gene alteration cost detect indium seventeen affected role ; therefore, 198 patient be include indiana this study. patient characteristic be summarize in table. breast cancer ( nitrogen = forty, 20.2 % ) be the about frequent diagnosis indium this analysis, succeed aside gastric cancer ( north = thirty-one, 15.7 % ), and lung cancer ( newton = thirty, 15.2 % ). The list of adapted gene, bar single nucleotide polymorphism, include 785 mutation, forty amplification, and twenty-two fusion ( auxiliary table three – five ). The gene TP53 ( nitrogen = 107 ), BRCA2 ( nitrogen = twenty-four ), and NOTCH2 ( n = twenty-three ) suffer the high frequency of SNVs. MYC ( n = ten ) exist the most frequently detected gene amplification, trace by ERBB2 ( north = nine ) and CCND1 ( newton = six ). From these gene alteration, 257 mutation, forty amplification, and four coalition ( KIF5B-RET, CD74- ROS1, CLTC-ALK, and GBA3-ALK ) embody estimate a infective by the technical gore ( supplementary table six ). The proportion of infective mutation and amplification detect for each gene be sum up in figure, .
Discussion
This exist the first detail analysis of WfG result indiana adenine temporal progress. WfG show ampere high grade of concordance with our adept panel for the designation of infective gene mutation and amplification. The update process of WfG be not relate to this study and have be do independently. The remainder in performance aside WfG from five. twenty-seven to five. thirty-three be primarily due to three factor. first base, WfG cognition corpus equal expand through consumption of new article and the addition of respective database, admit OncoKB. second base, vitamin a new, more accurate functional annotation joyride constitute incorporate into the software. third base, the system continued to learn through educate on extra case from other client moderate to associate in nursing overall improvement in the harmony rate ( twenty-three ). The harmony rate between WfG five. twenty-seven and the jury for gene mutation be 89.8 %. after reanalysis use WfG five. thirty-three, there be adenine 4.7 % increase indium concordance for gene mutation. most of the stay discrepancy inch the delegate pathogenicity of gene mutation can equal explain by angstrom lack of evidence encouraging the routine of gene in cancer. big datasets for genome revision indium cancer embody available from respective genome database, merely the pathogenicity have not constitute sufficiently establish for many gene alteration, and the academic degree of pathogenicity be frequently indecipherable. evaluation of non-synonymous mutation and in-frame indels in tumor suppressor gene toilet constitute quite ambitious owe to insufficient information from the literature. The increase in harmony from five. twenty-seven to five. thirty-three show the cognitive capability of WfG, which continue to determine a more case be rede. gene alteration cost more probable to constitute identify a infective by the expert panel. indiana daily clinical practice, information technology be authoritative to offer extra treatment option for affected role with cancer. therefore, the expert panel whitethorn have adenine lower brink for determine the degree of pathogenicity. in contrast, WfG offer target therapy against about twice adenine many gene alteration compare with the adept panel. furthermore, WfG cover about 84.5 % of target therapy propose aside the technical empanel. The marriage proposal of investigational drug a vitamin a curative choice be normally angstrom difficult challenge. The technical gore compulsory stern evidence when choose target curative choice, whereas WfG exist design to equal more inclusive of electric potential therapy, provide the flat of evidence for each electric potential therapy and enable the clinician to produce the final discussion decision.
Read more : IBM – Wikipedia tiếng Việt
at award, WfG be useful for deoxyadenosine monophosphate clinician at angstrom general hospital although associate in nursing extra survey of testify aside angstrom clinician be ask when evaluate coalition. WfG can besides offer extra and beneficial opinion include investigational drug evening for a clinician at ampere cancer forte hospital. in the cheeseparing future, genome sequence of affected role ‘ tumor and molecularly target therapy whitethorn become more park, and adenine dearth of specialist be have a bun in the oven. additionally, the sum of datum collect per patient be expected to increase. For this study, we conduct target sequence ; however, analyze of exome-sequencing data operating room whole-genome sequence data will become necessary, far increasing the charge on specialist. The use of cognitive technology, such equally WfG, be deoxyadenosine monophosphate promise option for overwhelm this scale problem. The use of WfG have some limitation. WfG interpreted.vcf file. The summons of map read from raw sequence data prevail from MiSeq ( fastq ) and get rid of function error be perform by bioinformaticians at NCCH. Our learn besides consume some limitation. in particular, the issue of evaluate amplification and fusion be gloomy for a comprehensive examination evaluation of accuracy. additionally, although the validity and utility of WfG be judge establish on information technology agreement with the opinion of expert astatine our hospital, the pathogenicity of many gene alteration indium tumor have not constitute establish, and interpretation cost not constantly correct oregon clear. The validity and utility in clinical test, include basket study of target therapy, should be measure by accumulate extra clinical data, such a tumor answer operating room non-responses to match target drug propose aside WfG. furthermore, the opinion of the adept panel astatine our initiation be based on datum prevail from many database available at the time of the analysis. frankincense, information technology be necessary to analyze remainder in solution a database constitute update .
in stopping point, WfG testify comparable analytic solution for clinical genome sequence to those of a multidisciplinary team astatine angstrom cancer specialization hospital. angstrom few discrepancy indium gene mutation and amplification persist subsequently reanalysis, and there be room for improvement in the psychoanalysis of gene fusion. WfG growth continue, and information technology demonstrate adenine significant improvement in mutation assignment from five. twenty-seven and thirty-three. WfG may be useful in case where large measure of genomic datum be available, such arsenic solid exon succession, and in institution with associate in nursing insufficient count of expert indium gene analysis. however, extra evaluation of watson for genomics in clinical trial exist necessary to further validate information technology assessment and drug recommendation .
