Runs - Precision Medicine 2018¶

aCSIROmedAll¶

Run ID: aCSIROmedAll
Participant: CSIROmed
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 88d177458e2b2aa89b0c2df5e1c6395a
Run description: We ran a BM25 ranking ith gene and disease boosting.

aCSIROmedMCB¶

Run ID: aCSIROmedMCB
Participant: CSIROmed
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 6fada5feba9d089abaa85e1406e3046b
Run description: We ran a BM25 ranking with gene and disease boosting and boosting cross citations and MeSH matching.

aCSIROmedNEG¶

Run ID: aCSIROmedNEG
Participant: CSIROmed
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: e175291afb8c9f9a8d465533b515478a
Run description: We ran a BM25 ranking with gene and disease boosting and negation detection and removal.

BB2_sq_nprf¶

Run ID: BB2_sq_nprf
Participant: Poznan
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 31a49803c425b1fc0267c3d735216c83
Run description: Dedicated method of creating a query. IR with BB2 model. No Query Expansion. Gene variant not taken.

BB2_vq_noprf¶

Run ID: BB2_vq_noprf
Participant: Poznan
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 29179f09570b7e07e0721169a30406da
Run description: Dedicated method of creating a query. IR with BB2 model. No Query Expansion methods applied.

BB2sqw2vprf¶

Run ID: BB2sqw2vprf
Participant: Poznan
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 0d4d8e89656931127473621ebc79fd3c
Run description: Dedicated method of creating a query. IR with BB2 model. Various Query Expansion methods applied. Gene variant not taken.

BB2vqw2vprf¶

Run ID: BB2vqw2vprf
Participant: Poznan
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 0a8d5fe4794adb0d23744d0fe4263aa8
Run description: Dedicated method of creating a query. IR with BB2 model. Various Query Expansion methods applied.

bool51¶

Run ID: bool51
Participant: Poznan
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: 418af7c56b2ac3baac70692f27f0fc64
Run description: Set of boolean queries with weights.

cbnuCT1¶

Run ID: cbnuCT1
Participant: cbnu
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: trials
MD5: 7bc45987db8bedf32c22d6e578238266
Run description: documents re-ranking using convolution neural network clustering

cbnuCT2¶

Run ID: cbnuCT2
Participant: cbnu
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: trials
MD5: 15986de15c144850574de2eaf7ac06ea
Run description: pseudo relevance feedback using convolution neural network clustering (treatment and age information priority)

cbnuCT3¶

Run ID: cbnuCT3
Participant: cbnu
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: trials
MD5: 789d3bbf2176bf9ea918f68c34f6c580
Run description: documents re-ranking using convolution neural network clustering (treatment and age information priority)

cbnuSA1¶

Run ID: cbnuSA1
Participant: cbnu
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: fc41fda9adce05e9e6ca31166a21257f
Run description: documents re-ranking using convolution neural network clustering

cbnuSA2¶

Run ID: cbnuSA2
Participant: cbnu
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: 8bcd805cdb8ad7ca95f000c9a49b1744
Run description: pseudo relevance feedback using convolution neural network clustering (treatment and age information priority)

cbnuSA3¶

Run ID: cbnuSA3
Participant: cbnu
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: 4f15a69004dc82f6ea7e3b14b04a1389
Run description: documents re-ranking using convolution neural network clustering (treatment and age information priority)

cCSIROmedAll¶

Run ID: cCSIROmedAll
Participant: CSIROmed
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 37770cd98b193d6fbfd5f5c7f9fddc9f
Run description: We ran a BM25 ranking with gene and disease boosting.

cCSIROmedHGB¶

Run ID: cCSIROmedHGB
Participant: CSIROmed
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 7426ae8913bb276891bf187f1d90f660
Run description: We ran a BM25 ranking with gene and disease boosting, cross-citation boosting, and matching of MeSH headings.

cCSIROmedNEG¶

Run ID: cCSIROmedNEG
Participant: CSIROmed
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: f37deac3ae650b037543962fd9d8e484
Run description: We ran a BM25 ranking with gene and disease boosting, negation detection and removal, and matching of MeSH headings.

cubicmnz¶

Run ID: cubicmnz
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 405de8b16bfd0bfe4c78f6f7d12bfaef
Run description: CombMNZ with two language models and bm25 retreival models and normalisation with cubic model.

cubicmnzAbs¶

Run ID: cubicmnzAbs
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 96a22dfeda3ce692313da955d533e756
Run description: CombMNZwith two language models and bm25 retrieval model and normalisation with cubic model.

cubicsumEW¶

Run ID: cubicsumEW
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: ff1418b29ccde39d0a5c3c1f2b5a7e7c
Run description: CombSum and cubic model and without normalisation.

cubicsumW¶

Run ID: cubicsumW
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 030bbd2fb12984edb5700f229842b19c
Run description: CombSum and cubic model with word embeddings.

cubicsumWAbs¶

Run ID: cubicsumWAbs
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 18ff197a83aa26d6a5895e7fbbcc3d25
Run description: CombSUM with wordEmbedding and cubic normalisation.

doc2vec_run¶

Run ID: doc2vec_run
Participant: ASU_Biomedical
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 65eaf30ddd3a6b9bc8149f654d28a273
Run description: This is an automatic run using doc2vec to get similar documents along with the scores.Resources used: 1. Python 3.5 2. Gensim Doc2Vec 3.5

doc2vec_run2¶

Run ID: doc2vec_run2
Participant: ASU_Biomedical
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: a3358f4638d51728291bb946f1a2cc3b
Run description: Supervised approach to predict rank of similar documents.

ECNU_C_Run1¶

Run ID: ECNU_C_Run1
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 64596c1ec1bc85700877e30e38b0c105
Run description: The result retreived on entire clinic-trials dataset by terrier platform with BM25 model and pseudo relevance feedback.

ECNU_C_Run2¶

Run ID: ECNU_C_Run2
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: be0b8e1db1fbba56366b1f2a0acefe5e
Run description: Combine the PRF and BM25 to retreived on entire clinic-trials dataset.

ECNU_C_Run3¶

Run ID: ECNU_C_Run3
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 30247a922ac1fd470ffc7f8b56ddad1c
Run description: Add the descriptions of diseases as expansion of each topic, which are crawled from the NCBI Mesh, then do the same as RUN1.

ECNU_C_Run4¶

Run ID: ECNU_C_Run4
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: c8cb814de10c899342d7c044c129f0bf
Run description: Combination of RUN 1 to 3.

ECNU_C_Run5¶

Run ID: ECNU_C_Run5
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: e05ea92b69a143b7ab01540ce1d4d820
Run description: Based on RUN4, delete the trials unmatched the constrains such as ages.

ECNU_S_Run1¶

Run ID: ECNU_S_Run1
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: dc5019e24efb5d014f018d04109a71c0
Run description: The result retreived on entire Medline and extra-abstract dataset by terrier platform with BM25 model and pseudo relevance feedback.

ECNU_S_Run2¶

Run ID: ECNU_S_Run2
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 9b65af46236d5f80098942e1b44ddbc8
Run description: For each topic, rerank fisrt 1000 documents retrieved in RUN1 by using LDA-Model to calculate the average similarity between all terms of each topic and related documents.

ECNU_S_Run3¶

Run ID: ECNU_S_Run3
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 2116eea09b2abe7443a17f7ecb719581
Run description: For each topic, rerank fisrt 1000 documents retrieved in RUN1 by using LDA-Model to calculate the average similarity between all terms of each topic and related documents. The terms are included high-frequency words from the frst 20 NCBI-PubMed abstracts which was related to the original topic.

ECNU_S_Run4¶

Run ID: ECNU_S_Run4
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 43e8add69c36c258b029bb82a5f79561
Run description: For each topic, rerank fisrt 2000 documents retrieved in RUN1 by using LDA-Model to calculate the average similarity between all terms of each topic and related documents, and finally choose the fisrt 1000 documents as results.

ECNU_S_Run5¶

Run ID: ECNU_S_Run5
Participant: ECNUica
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 86d61c4afa0f6819ca061c226796c0ab
Run description: For each topic, rerank fisrt 2000 documents retrieved in RUN1 by using LDA-Model to calculate the average similarity between all terms of each topic and related documents, and finally choose the fisrt 1000 documents as results. The terms are included high-frequency words from the frst 20 NCBI-PubMed abstracts which was related to the original topic.

elastic_run¶

Run ID: elastic_run
Participant: ASU_Biomedical
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: 9572e9cb82b98c6b7a3f07782d4d82b2
Run description: This is a elastic search based automated run based on Python. We have used following frameworks for this automated run: 1. Python 3.5 2. Elastic Search 6.3

hpictall¶

Run ID: hpictall
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: trials
MD5: 29401e5c728bb7297e84f2a5d4991aa9
Run description: Same as "hpictboost", but with gene descriptions from NCBI gene list.

hpictbase¶

Run ID: hpictbase
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: trials
MD5: 7ce2743905c09b91d4dee5f99e8b5d99
Run description: Baseline run. Matching only disease and gene "as is". Boosts as in "hpictboost" run. Excluding non-melanoma.

hpictboost¶

Run ID: hpictboost
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: bec77448844481d107d204d7210f5369
Run description: Matching age, sex, the disease preferred term and its synonyms (from SNOMED CT, MeSH, and ICD), gene synonyms (from NCBI gene list) and its family, as well as solid tumors. Boosting cancer and genetics keywords, as well as a list of positive handcrafted boosters. Excluding non-melanoma.

hpictcommon¶

Run ID: hpictcommon
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: trials
MD5: 5520a83929f55447261137bdec3b2977
Run description: Same as "hpictboost", but without solid tumor expansion and gene family, in order to have a comparable run with Biomedical Articles.

hpictphrase¶

Run ID: hpictphrase
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: trials
MD5: 1cc159302e421199a21fcef6f4c6eece
Run description: Same as "hpictboost, but with exact matching on diseases.

hpipubbase¶

Run ID: hpipubbase
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: bd739d133dc71743158ce5a211309df6
Run description: Baseline run. Matching only disease and gene "as is". Boosts as in "hpipubboost" run, but without the ML classifier. Excluding non-melanoma.

hpipubboost¶

Run ID: hpipubboost
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 14e64612a72a39f31bf8ed2e9e360ef1
Run description: Matching the disease preferred term and its synonyms (from SNOMED CT, MeSH, and ICD), the gene description and its synonyms (from NCBI gene list). Boosting AACR/ASCO abstracts, chemotherapy suffixes, cancer and genetics keywords, positive and negative handcrafted boosters, as well as a positive prediction from a ML classifier trained on the 2017 gold standard. Excluding non-melanoma.

hpipubclass¶

Run ID: hpipubclass
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: 17152104525c9e3485ab2292bbb19d5e
Run description: Matching the disease preferred term and its synonyms (from SNOMED CT, MeSH, and ICD), the gene description and its synonyms (from NCBI gene list). Boosting chemotherapy suffixes, cancer and genetics keywords, positive and negative handcrafted boosters, as well as a highly boosted PM/Not PM MALLET MaxEnt classifier with SecondString TFIDF estimates, BANNER tagger gene counter and MeSH headings from the Medline XML. Excluding non-melanoma.

hpipubcommon¶

Run ID: hpipubcommon
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: 0ace8c5d18e310e06c5232d031f1778e
Run description: Matching the disease preferred term and its synonyms (from SNOMED CT, MeSH, and ICD), gene synonyms (from NCBI gene list). Boosting chemotherapy suffixes, cancer and genetics keywords, positive and negative handcrafted boosters. Excluding non-melanoma.

hpipubnone¶

Run ID: hpipubnone
Participant: hpi-dhc
Track: Precision Medicine
Year: 2018
Submission: 8/6/2018
Type: automatic
Task: abstracts
MD5: 05422ebf367ca45903327c7b04e6172c
Run description: Matching the disease preferred term and its synonyms (from SNOMED CT, MeSH, and ICD), the gene description and its synonyms (from NCBI gene list). Boosting chemotherapy suffixes, cancer and genetics keywords, positive and negative handcrafted boosters. Excluding non-melanoma.

IKM_trail_1¶

Run ID: IKM_trail_1
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: 3c2850c292e646f136a5fd2463fe35c8
Run description: Normal(include Exclusion)

IKM_trail_2¶

Run ID: IKM_trail_2
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: 0d52fac6fc5b9dd8f192d053a543f4fb
Run description: Normal + Minus Exclusion

IKM_trail_3¶

Run ID: IKM_trail_3
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: a666ec9bb5205ac3ff8f7a06a0acdb93
Run description: Normal + Include Only Disease

IKM_trail_4¶

Run ID: IKM_trail_4
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: eebd8c390e121792f8c8ad0c9e7d6d3e
Run description: Normal + Include Only Gene

IKM_trail_5¶

Run ID: IKM_trail_5
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: 86e8f8fb6888c5df94c8676d74a8c219
Run description: Normal - Low Relationship Tags

IKMLAB_1¶

Run ID: IKMLAB_1
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: 2a629dcdc294680845fd7ce8e7825992
Run description: Pubtator + Chemical

IKMLAB_2¶

Run ID: IKMLAB_2
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: d2f30980e973e5cc2cd729f9db63f189
Run description: Pubtator + No Chemical

IKMLAB_3¶

Run ID: IKMLAB_3
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: 56b6948f3bcefe30908345c3b08100ed
Run description: Extra + No Species and Chemical

IKMLAB_4¶

Run ID: IKMLAB_4
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: 1cfaba33c358e33555c2fddb0c87654f
Run description: Extra + Species and Chemical

IKMLAB_5¶

Run ID: IKMLAB_5
Participant: IKMLAB
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: fa13208c69331c1e40cef580e31bf270
Run description: Using a neural network to get the answer.

imi_mug_abs1¶

Run ID: imi_mug_abs1
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: abstracts
MD5: c0525c8fd5e93537ef9001b504119f9f
Run description: Baseline query with MeSH age groups and grid-search optimized parameters.

imi_mug_abs2¶

Run ID: imi_mug_abs2
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 8b9db5cfa78d78c44feabc94e73f8f66
Run description: Baseline query with MeSH age groups and grid-search optimized parameters. Boost English and extra topics.

imi_mug_abs3¶

Run ID: imi_mug_abs3
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 59443ed9ea1e33511544bc0372924662
Run description: Baseline query with MeSH age groups and grid-search optimized parameters. Boost English and extra topics. Search age group and sex in title, abstract, and MeSH.

imi_mug_abs4¶

Run ID: imi_mug_abs4
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 19823f0294c4ea1feb3e4debd44e1b01
Run description: Baseline query with MeSH age groups and grid-search optimized parameters. Boost English and extra topics. Search age group and sex in title, abstract, and MeSH. Abstract should exist.

imi_mug_abs5¶

Run ID: imi_mug_abs5
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 95f41a33e4a808192a8b193e27e788b1
Run description: Baseline query with MeSH age groups and grid-search optimized parameters. Boost English and extra topics. Abstract should exist.

imi_mug_ct1¶

Run ID: imi_mug_ct1
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: trials
MD5: 9f7ff828d4d51851040fc769db03c96e
Run description: Baseline query with bi and tri-gram search, plus grid-search optimized parameters.

imi_mug_ct2¶

Run ID: imi_mug_ct2
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 8d000cfcfe8a92defded6b156ec84bc7
Run description: Baseline query with bi and tri-gram search, plus grid-search optimized parameters. Disease must not match exclusion criteria.

imi_mug_ct3¶

Run ID: imi_mug_ct3
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: facc917835767ac0bec7af310f26d050
Run description: Baseline query with bi and tri-gram search, plus grid-search optimized parameters. All should (relaxed).

imi_mug_ct4¶

Run ID: imi_mug_ct4
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: ac7174119160dc2dd32f29569d9223da
Run description: Baseline query with bi and tri-gram search, weighed, plus grid-search optimized parameters.

imi_mug_ct5¶

Run ID: imi_mug_ct5
Participant: imi_mug
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: d268ba2490cab12992f50b1e9d429109
Run description: Baseline query with bi and tri-gram search, weighed, plus grid-search optimized parameters.

IMS_NO_PRF¶

Run ID: IMS_NO_PRF
Participant: ims_unipd
Track: Precision Medicine
Year: 2018
Submission: 7/30/2018
Type: automatic
Task: trials
MD5: 1cdd780ef65946afcbdd7396b5b3d4c1
Run description: Term-concept based retrieval using concepts within a knowledge-base (UMLS Metathesaurus) to expand the original query. The scoring function is the BM25F (Okapi version).

IMS_PRF¶

Run ID: IMS_PRF
Participant: ims_unipd
Track: Precision Medicine
Year: 2018
Submission: 7/30/2018
Type: automatic
Task: trials
MD5: c91e070eb155ef0b7aaf0e8f7a448330
Run description: term-concept based retrieval with Pseudo Relevance Feedback to expand query with concepts-terms that are related to the query terms. The selected concepts-terms are those that have a relation within the UMLS Metathesaurus with the concepts in the original query. The scoring function used is the BM25F (Okapi).

IMS_TERM¶

Run ID: IMS_TERM
Participant: ims_unipd
Track: Precision Medicine
Year: 2018
Submission: 7/30/2018
Type: automatic
Task: trials
MD5: 8b4076e476c94e92db65de8388560a51
Run description: Term-based retrieval using original terms contained within the query. The scoring function is the BM25F (Okapi version).

irit_prf_cli¶

Run ID: irit_prf_cli
Participant: IRIT
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: trials
MD5: 25eadea5338ad5fd9d3c85af8875fcaa
Run description: Classical PRF with Indri, nDoc=5 nTerm=10

KL18AbsFuse¶

Run ID: KL18AbsFuse
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: a735a3598b0a54affdb35fc74315b5a7
Run description: Fusion runs: - Fusion of two different expansion methods and baseline run

KL18absHY¶

Run ID: KL18absHY
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 57c67989193135a53160b3605e22cb71
Run description: Expansion with NCIT as KB

KL18absWV¶

Run ID: KL18absWV
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 8fad1acf30726557de31749674e721ac
Run description: - Pooling w2vec expansion run for diversity

KL18TrialBF¶

Run ID: KL18TrialBF
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 9c5ee3a433b02c3858a7c65d52cc5b5d
Run description: Different baseline with filtering

KL18TrialWV¶

Run ID: KL18TrialWV
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: a0f2dab422760c26faa4627d25ca59f6
Run description: Word2vec expansion run for diversity

KL18TriFuse¶

Run ID: KL18TriFuse
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 9e563c7358b6ea571fa68c336a722517
Run description: Fusion run for Trials .

KL18TriHY¶

Run ID: KL18TriHY
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: ea1731903a04885691038be66001a2c1
Run description: Expansion using NCIT

KLPM18T1Bl¶

Run ID: KLPM18T1Bl
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: 3ba8680e2d5068e6530cde0c9f775047
Run description: _ baseline with filtering on demographic data,

KLPM18T2Bl¶

Run ID: KLPM18T2Bl
Participant: KlickLabs
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: fa897b1aff53beecc3754d7722301f4a
Run description: _ baseline Abstracts with filtering on demographic data,

m_trial1¶

Run ID: m_trial1
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 145a34e18190c69ce1785ed4ab0493dd
Run description: method fusion for trial

mayoctcomp¶

Run ID: mayoctcomp
Participant: MayoNLPTeam
Track: Precision Medicine
Year: 2018
Submission: 8/1/2018
Type: automatic
Task: trials
MD5: 1c2b3bf7f45acf19278421d5a49d2b4a
Run description: Comprehensive method combining entity recognition and learning to rank

mayoctscreat¶

Run ID: mayoctscreat
Participant: MayoNLPTeam
Track: Precision Medicine
Year: 2018
Submission: 8/1/2018
Type: automatic
Task: trials
MD5: db8ffe512ef3a3090e3dc01ee76bc974
Run description: using ctakes results for searching disease and conditions

mayoctsimp¶

Run ID: mayoctsimp
Participant: MayoNLPTeam
Track: Precision Medicine
Year: 2018
Submission: 8/1/2018
Type: automatic
Task: trials
MD5: 7ff1e76c3170569df310ca62826f58bc
Run description: Baseline method using keyword search

mayomedcomp¶

Run ID: mayomedcomp
Participant: MayoNLPTeam
Track: Precision Medicine
Year: 2018
Submission: 8/1/2018
Type: automatic
Task: abstracts
MD5: a3210c59d376e63900554845386c2cb5
Run description: comprehensive retrieval model using learning to rank

mayomedcreat¶

Run ID: mayomedcreat
Participant: MayoNLPTeam
Track: Precision Medicine
Year: 2018
Submission: 8/1/2018
Type: automatic
Task: abstracts
MD5: 121d92d7f70399cfb099d55fd82d0fe5
Run description: using ctakes results for searching disease and conditions

mayomedsimp¶

Run ID: mayomedsimp
Participant: MayoNLPTeam
Track: Precision Medicine
Year: 2018
Submission: 8/1/2018
Type: automatic
Task: abstracts
MD5: bbf9fdad63329172348929e5d717e32c
Run description: Simple baseline method using keyword and MeSH term matching.

mayopubtator¶

Run ID: mayopubtator
Participant: MayoNLPTeam
Track: Precision Medicine
Year: 2018
Submission: 8/1/2018
Type: automatic
Task: abstracts
MD5: 4c3f6f532b988db877a31b5b81aad3a6
Run description: using pubtator developed by NIH for searching diseases and genes

MedIER_sa11¶

Run ID: MedIER_sa11
Participant: MedIER
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 4d1d38426844711bb40e0587d0f1b813
Run description: BM25, appended results from multiple queries; Resources: Gene ontology, PubTator

MedIER_sa12¶

Run ID: MedIER_sa12
Participant: MedIER
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: f9ac0df3ffe92d1f620b512a60e5c175
Run description: LM, appended results from multiple queries; Resources: Gene ontology, PubTator

MedIER_sa13¶

Run ID: MedIER_sa13
Participant: MedIER
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 80f640fc0264fed34ef4daafca892163
Run description: RRF Fusion on runs MedIER_sa11 and MedIER_sa12

MedIER_sa14¶

Run ID: MedIER_sa14
Participant: MedIER
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 5aceb04004c2c733898affefdd42b88a
Run description: BM25, appended results from multiple queries; re-ranked using pseudo-relevance classifier; Resources: Gene ontology, PubTator

MedIER_sa15¶

Run ID: MedIER_sa15
Participant: MedIER
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 53cfb2584add309179e9d4935018471d
Run description: LM, appended results from multiple queries; re-ranked using pseudo-relevance classifier; Resources: Gene ontology, PubTator

method_fu¶

Run ID: method_fu
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: c75b39da4ab0b69600a40b9439a87755
Run description: method fusion

minfolabBA¶

Run ID: minfolabBA
Participant: InfoLabPM
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: abstracts
MD5: 52766488dac18c593ab9fbb80d16c864
Run description: The matching model used: BB2 Direct use of disease, gene and gender

minfolabBC¶

Run ID: minfolabBC
Participant: InfoLabPM
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: abstracts
MD5: 43f597b4ee4aef2b775c0e7481d6bf83
Run description: The matching model used: BB2 Expanded disease term to the most preferred atom (via UMLS REST API)

minfolabBD¶

Run ID: minfolabBD
Participant: InfoLabPM
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: abstracts
MD5: 1237a593765b2afb1422193fb7b9724a
Run description: The matching model used: BB2 Expanded disease term to the most preferred atom (via UMLS REST API) Expanded Genes terms (via Ensembl REST API)

minfolabTH¶

Run ID: minfolabTH
Participant: InfoLabPM
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: abstracts
MD5: b60f7699e100bc3e0ba256d145961826
Run description: The matching model used: BB2 Expanded with human and neoplasm keywords https://www.nlm.nih.gov/bsd/pubmed_subsets/cancer_strategy.html

mnz¶

Run ID: mnz
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 09de909edc63930fae39d6bd1f23e674
Run description: CombMNZ with two language models and bm25 retreival models.

mnzAbs¶

Run ID: mnzAbs
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: f259d4328d9088bc02d50e807bcd3d89
Run description: CombMNZwith two language models and bm25 retrieval model.

MSIIP_BASE¶

Run ID: MSIIP_BASE
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 36cde70ea4036f284a7a38e5ad5ba63f
Run description: A baseline system.

MSIIP_PBAH¶

Run ID: MSIIP_PBAH
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 03bde58a7d4008a040642667dfc422d6
Run description: Using our PM classifier with a high weight upon all the articles.

MSIIP_PBH¶

Run ID: MSIIP_PBH
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: dcd275be734c5197fae38057a04f2beb
Run description: Using our PM classifier with a high weight upon all the articles except for the top 10.

MSIIP_PBL¶

Run ID: MSIIP_PBL
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 58e9c2afa3eeea45e8cb9d8d6410bdec
Run description: Using our PM classifier with a low weight upon all the articles except for the top 10.

MSIIP_PBPK¶

Run ID: MSIIP_PBPK
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 2e473bf980947588976c86d32081415a
Run description: Using our PM classifier along with the PM classification function of our baseline system in the way of disjunctive.

MSIIP_TRIAL1¶

Run ID: MSIIP_TRIAL1
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 191eaa052119d6ebc3b9df5ae09e7dc5
Run description: msiip_trials3, without condidering 'study_type', 'intervention_type' and 'primary_purpose' fields.

MSIIP_TRIAL2¶

Run ID: MSIIP_TRIAL2
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 808d55b38362bc6e44feb21a814cfb46
Run description: msiip_trials5, not containing query_expansion, without condidering 'study_type', 'intervention_type' and 'primary_purpose' fields.

MSIIP_TRIAL3¶

Run ID: MSIIP_TRIAL3
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 3ecd03dd735d8a5f279bbf07b4249ede
Run description: containing all our methods, with some differential methods for some topics involving immonotherapy

MSIIP_TRIAL4¶

Run ID: MSIIP_TRIAL4
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 364abe13393ffb2ccb41171b3014c15d
Run description: msiip_trials5, not containing query_expansion.

MSIIP_TRIAL5¶

Run ID: MSIIP_TRIAL5
Participant: Cat_Garfield
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 59a5940ba1b25bda1ce3aecfc668b2ef
Run description: containing all our methods, all topics share the same methods

NS_PM_1¶

Run ID: NS_PM_1
Participant: NOVASearch
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: da978f3c3eeef04b10d9422fde30244b
Run description: Base search using Lucene. Re ranking based on three criteria: - Purpose, intervention and study type - Gene type matching the inclusion criteria, expanded using the HGNC database. - Patient condition matching trial condition, expanded using the MeSH and SNOMed databases.

NS_PM_2¶

Run ID: NS_PM_2
Participant: NOVASearch
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: d6ed6f9c1472937cadae6811a32417fa
Run description: Base search using Lucene. Re ranking based on three criteria: - Purpose, intervention and study type - Gene type matching the inclusion criteria, expanded using the HGNC database.

NS_PM_3¶

Run ID: NS_PM_3
Participant: NOVASearch
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: 69aece298477376063e4e1b7546c96d7
Run description: LETOR using LambdaMart. Features based on multiple retrieval functions (BM25, TF-IDF, Language Models) and features based on to NS_PM_1 and 2 filters

NS_PM_4¶

Run ID: NS_PM_4
Participant: NOVASearch
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: 83c4f7db1e0010e892b477c8edf1321f
Run description: LETOR using AdaRank. Features based on multiple retrieval functions (BM25, TF-IDF, Language Models) and features based on to NS_PM_1 and 2 filters

NS_PM_5¶

Run ID: NS_PM_5
Participant: NOVASearch
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: e33b6095bf3f4360668dabd013d698c0
Run description: LETOR using Coordinate Ascend. Features based on multiple retrieval functions (BM25, TF-IDF, Language Models) and features based on to NS_PM_1 and 2 filters.

para_fusion¶

Run ID: para_fusion
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: cfea70527a6da887bd43afe4350e872c
Run description: parameter fusion

para_trial¶

Run ID: para_trial
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: a4a8a754bc23da6ec89d52b4d908664e
Run description: parameter fusion

PM_IBI_run1¶

Run ID: PM_IBI_run1
Participant: PM_IBI
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: 1319df6a4695e4bbf5864d255e3752ad
Run description: PM_IBI_run1 description: For each topic, we automatically expand the terms describing the related disease(s), gene(s) and variant(s) by relying on the UMLS meta-thesaurus and semantic network and on dbSNP. In particular for each gene and disease term we automatically gathered from UMLS: (i) all its synonym terms; (ii) all the terms describing more specific concepts (hyponyms) by means of the UMLS semantic network (traversed up to a maximum depth of three levels). When variant(s) are present, we retrieved its reference SNP ID number (when possible). The automatically expanded lists of terms - disease(s), gene(s) and variant(s) - are jointly used to properly structure a boolean query to an ElasticSearch document indexer where the whole set of PubMed abstracts provided to the challenge participants has been previously loaded. The top-1000 results retrieved by Elasticsearch are re-ranked by taking into account the following six relevance indicators: (i) for each gene and disease, the number of mentions of the original term and its synonyms (from UMLS); (ii) for each gene and disease, the number of mentions of terms describing more specific concepts (hyponyms - from UMLS); (iii) in case the gene name is present and the related variant is mentioned too, the number of mentions of the variant; (iv) the number of mentions of chemicals (MeSH compounds and string matches from a list of compound names derived from UMLS); (v) the presence, in a customized version of the bio-marker response database ResMarker (http://resmarkerdb.org), of a validated relation among the genes, diseases, variants and chemicals mentioned in the abstract; (vi) the year of publication of the abstract. In the PM_IBI_run1, we generated the final list of ranked results for each topic by considering sequentially the following sub-groups of the abstracts returned by Elasticsearch (each sub-group is in turns internally ordered by relying on the previous set of six relevance indicators): A) abstracts with at least one mention of a disease (exact or more specific term), a gene (exact or more specific term) and a variant in the title B) abstracts with at least one mention of a disease (exact or more specific term) and a gene (exact or more specific term) in the title (no variant mentions) C) abstracts with at least one mention of a disease (exact or more specific term), a gene (exact or more specific term) and a variant considering both title and abstract D) abstracts with at least one mention of a disease (exact or more specific term) and a gene (exact or more specific term) considering both title and abstract (no variant mentions) E) abstracts with one mention of a disease (exact or more specific term) or (XOR) one mention of a gene (exact or more specific term) together with one or more variant mentions, considering both title and abstract F) abstracts with one mention of a disease (exact or more specific term) or (XOR) one mention of a gene (exact or more specific term) considering both title and abstract (no variant mentions)

PM_IBI_run2¶

Run ID: PM_IBI_run2
Participant: PM_IBI
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: b665910ba02810038130f68cf7422e07
Run description: PM_IBI_run2 description: For each topic, we automatically expand the terms describing the related disease(s), gene(s) and variant(s) by relying on the UMLS meta-thesaurus and semantic network and on dbSNP. In particular for each gene and disease term we automatically gathered from UMLS: (i) all its synonym terms; (ii) all the terms describing more specific concepts (hyponyms) by means of the UMLS semantic network (traversed up to a maximum depth of three levels). When variant(s) are present, we retrieved its reference SNP ID number (when possible). The automatically expanded lists of terms - disease(s), gene(s) and variant(s) - are jointly used to properly structure a boolean query to an ElasticSearch document indexer where the whole set of PubMed abstracts provided to the challenge participants has been previously loaded. The top-1000 results retrieved by Elasticsearch are re-ranked by taking into account the following six relevance indicators: (i) for each gene and disease, the number of mentions of the original term and its synonyms (from UMLS); (ii) for each gene and disease, the number of mentions of terms describing more specific concepts (hyponyms - from UMLS); (iii) in case the gene name is present and the related variant is mentioned too, the number of mentions of the variant; (iv) the number of mentions of chemicals (MeSH compounds and string matches from a list of compound names derived from UMLS); (v) the presence, in a customized version of the bio-marker response database ResMarker (http://resmarkerdb.org), of a validated relation among the genes, diseases, variants and chemicals mentioned in the abstract; (vi) the year of publication of the abstract. In the PM_IBI_run2, we generated the final list of ranked results for each topic by considering sequentially the following sub-groups of the abstracts returned by Elasticsearch (each sub-group is in turns internally ordered by relying on the previous set of six relevance indicators): A) abstracts with at least one mention of a disease (exact or more specific term), a gene (exact or more specific term) and a variant, considering both title and abstract B) abstracts with at least one mention of a disease (exact or more specific term) and a gene (exact or more specific term), considering both title and abstract (no variant mentions) C) abstracts with one mention of a disease (exact or more specific term) or (XOR) one mention of a gene (exact or more specific term) eventually with one or more variant mentions, considering both title and abstract (In our other run, PM_IBI_run1, we always give highest rank if a paper has gene, disease and variant mentions in the titile. Here title and abstract are jointly considered as texts that give the same relevance to mentions that occur in.)

PM_IBI_run3¶

Run ID: PM_IBI_run3
Participant: PM_IBI
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: 0c77bab9f9125351a14745f38281e15d
Run description: PM_IBI_run3 description: Contrary to the other two runs of the PM_IBI team (run1 and run2), this run, PM_IBI_run3, exploits the demographic match between the topic and the abstract to better rank search results. For each topic, we automatically expand the terms describing the related disease(s), gene(s) and variant(s) by relying on the UMLS meta-thesaurus and semantic network and on dbSNP. In particular for each gene and disease term we automatically gathered from UMLS: (i) all its synonym terms; (ii) all the terms describing more specific concepts (hyponyms) by means of the UMLS semantic network (traversed up to a maximum depth of three levels). When variant(s) are present, we retrieved its reference SNP ID number (when possible). The automatically expanded lists of terms - disease(s), gene(s) and variant(s) - are jointly used to properly structure a boolean query to an ElasticSearch document indexer where the whole set of PubMed abstracts provided to the challenge participants has been previously loaded. The top-1000 results retrieved by Elasticsearch are re-ranked by taking into account the following six relevance indicators: (i) for each gene and disease, the number of mentions of the original term and its synonyms (from UMLS); (ii) for each gene and disease, the number of mentions of terms describing more specific concepts (hyponyms - from UMLS); (iii) in case the gene name is present and the related variant is mentioned too, the number of mentions of the variant; (iv) the number of mentions of chemicals (MeSH compounds and string matches from a list of compound names derived from UMLS); (v) the presence, in a customized version of the bio-marker response database ResMarker (http://resmarkerdb.org), of a validated relation among the genes, diseases, variants and chemicals mentioned in the abstract; (vi) the year of publication of the abstract. In the PM_IBI_run3, we extract from each abstract returned by Elasticsearch demographic information (age-ranges of patients), where available. Thus, given the age of the patient specified by the topic description, we can identify all the abstracts in which the age of the patient is included in one or more age-ranges. As a consequence, we divide the results retrieved by the Elasticsearch query in the two following groups: 1) abstracts with one or more age-ranges including the age of the patient 2) abstract with no matching age-range or without age-range specified. Internally we order the abstracts of each of these two groups by considering sequentially the following sub-groups (each sub-group is in turns internally ordered by relying on the previous set of six relevance indicators): A) abstracts with at least one mention of a disease (exact or more specific term), a gene (exact or more specific term) and a variant in the title B) abstracts with at least one mention of a disease (exact or more specific term) and a gene (exact or more specific term) in the title (no variant mentions) C) abstracts with at least one mention of a disease (exact or more specific term), a gene (exact or more specific term) and a variant considering both title and abstract D) abstracts with at least one mention of a disease (exact or more specific term) and a gene (exact or more specific term) considering both title and abstract (no variant mentions) E) abstracts with one mention of a disease (exact or more specific term) or (XOR) one mention of a gene (exact or more specific term) together with one or more variant mentions, considering both title and abstract F) abstracts with one mention of a disease (exact or more specific term) or (XOR) one mention of a gene (exact or more specific term) considering both title and abstract (no variant mentions)

raw_medline¶

Run ID: raw_medline
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: 555abb90ee1ca0519756fca38a7c4de9
Run description: Naive method for task 1.

raw_trials¶

Run ID: raw_trials
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: trials
MD5: bc082d4d066ea6d93699383347015374
Run description: Naive method from clinical trials.

rbf¶

Run ID: rbf
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 8aa244dab315e94c29cdee8555e8c842
Run description: feedback

rfb_trial1¶

Run ID: rfb_trial1
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: d6d1cab45f4a6537aa7d0d523741a2cc
Run description: feedback

RSA_DSC_CT_1¶

Run ID: RSA_DSC_CT_1
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: trials
MD5: 8580f5ef1b7dcbc226173db9ac875d6d
Run description: The following stack of queries was used: 1. pr_pr_ex 2. prpr_df 3. sepr_ex 4. sepr_df 5. prse_ex 6. pr_se_df ... A GATE annotation pipeline was used in the topics' processing

RSA_DSC_CT_2¶

Run ID: RSA_DSC_CT_2
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: trials
MD5: e53eb48958979e4e51c943fd0763cc7e
Run description: The following stack of queries was used: 1. pr_pr_ex 2. prpr_df 3. sese_ex 4. sese_df 5. sepr_ex 6. se_pr_df ... A GATE annotation pipeline was used in the topics' processing

RSA_DSC_CT_3¶

Run ID: RSA_DSC_CT_3
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: trials
MD5: 7ed5dfdadbcc22a721ee8c9dfed6b87c
Run description: Used the following stack of queries: 1. pr_pr_ex 2. prpr_df 3. prse_ex 4. prse_df 5. sese_ex 6. se_se_df ... A GATE annotation pipeline was used in the topics' processing

RSA_DSC_CT_4¶

Run ID: RSA_DSC_CT_4
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: trials
MD5: e0a574972d878d8725898dc49280c859
Run description: Uses the following stack of queries: 1. pr_pr_ex 2. prse_ex 3. sese_ex 4. pr_pr_df ... A GATE annotation pipeline was used in the topics' processing.

RSA_DSC_CT_5¶

Run ID: RSA_DSC_CT_5
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: trials
MD5: ebf5635735a6b44349131f8f7d2dc0f5
Run description: The following rank of queries was used: 1. pr_pr_ex 2. prse_ex 3. prpr_df 4. se_se_ex .... A GATE annotationa pipeline was used in the topics' processing

RSA_DSC_LA_1¶

Run ID: RSA_DSC_LA_1
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: abstracts
MD5: c4321e77c59e672874d2e31eaeb53864
Run description: Used the following stack of queries: 1. pr_pr_ex 2. prpr_df 3. prse_ex 4. prse_df 5. sese_ex 6. se_se_df ... A GATE annotation pipeline was used in the topics' processing

RSA_DSC_LA_2¶

Run ID: RSA_DSC_LA_2
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: abstracts
MD5: 563a46b37ff9fbfdd3bbc17e7c8354d8
Run description: Used the following stack of queries: 1. pr_pr_ex 2. prse_ex 3. sese_ex 4. prpr_df 5. sepr_ex 6. pr_se_df ... A GATE annotation pipeline was used in the topics' processing

RSA_DSC_LA_3¶

Run ID: RSA_DSC_LA_3
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: abstracts
MD5: 024532125f0650d09531a9fcf6d622ca
Run description: Used the following stack of queries: 1. pr_pr_ex 2. prpr_df 3. prse_ex 4. sese_ex 5. sese_df 6. pr_se_df ... A GATE annotation pipeline was used in the topics' processing

RSA_DSC_LA_4¶

Run ID: RSA_DSC_LA_4
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: abstracts
MD5: 5d8463864b8ef94383703ddaef285f60
Run description: Used the following stack of queries: 1. pr_pr_ex 2. prse_ex 3. prpr_df 4. sese_ex 5. se*_se_df ... A GATE annotation pipeline was used in the topics' processing

RSA_DSC_LA_5¶

Run ID: RSA_DSC_LA_5
Participant: RSA_DSC
Track: Precision Medicine
Year: 2018
Submission: 7/27/2018
Type: automatic
Task: abstracts
MD5: 2ff3d0ac9dc91ab93882292832f07196
Run description: Used the following stack of queries: 1. pr_pr_ex 2. prse_ex 3. prpr_df 4. prse_df 5. sese_ex 6. se_se_df ... A GATE annotation pipeline was used in the topics' processing

SIBTMct1¶

Run ID: SIBTMct1
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: 2e4e1e1d98303bbca4a464b9117a79b0
Run description: Diseases are mapped and CTs are filtered by topics thanks to NCI thesaurus in : - conditions, mesh conditions, and keywords fields - along with titles and summary Genes are used in a standard search engine, Terrier, BM25. Demographic features are filtered. Boost are granted according to phases, primary purposes, and study types. Boosting values were computed with last year qrel distribution.

SIBTMct2¶

Run ID: SIBTMct2
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: 2fc07eaf9e62f3c8a90f395bb6bb9e0f
Run description: Diseases are mapped and CTs are filtered by topics thanks to NCI thesaurus in : - conditions, mesh conditions, and keywords fields Genes are used in a standard search engine, Terrier, BM25. Diseases and mutation keywords are also used in query. Demographic features are filtered. Boost are granted according to phases, primary purposes, and study types. Boosting values were computed with last year qrel distribution.

SIBTMct3¶

Run ID: SIBTMct3
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: 99700f6febda8807adfa7f27fb1f6187
Run description: Diseases are mapped and CTs are filtered by topics thanks to NCI thesaurus in : - conditions, mesh conditions, and keywords fields Genes are used in a standard search engine, Terrier, BM25. Demographic features are filtered. Boost are granted according to phases, primary purposes, and study types. Boosting values were computed with last year qrel distribution.

SIBTMct4¶

Run ID: SIBTMct4
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: efcacec6c9ce90ee33dd7469566d423a
Run description: Diseases are mapped and CTs are filtered by topics thanks to NCI thesaurus in : - conditions, mesh conditions, and keywords fields - along with titles and summary Genes are used in a standard search engine, Terrier, BM25. Diseases and mutation keywords are also used in query. Demographic features are filtered. Boost are granted according to phases, primary purposes, and study types. Boosting values were computed with last year qrel distribution.

SIBTMlit1¶

Run ID: SIBTMlit1
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 7/23/2018
Type: automatic
Task: abstracts
MD5: a6c9e0ca4a7d67e1e0eb1824fbe4f23e
Run description: Run 1 is our baseline run. Four queries, enabling constraint relaxing (disease+gene+variant; disease+gene; disease+variant; gene+variant) are sent to our search engine. Results are re-ranked based on drug density. Drug density is based on a list of drugs provided by DrugBank, coupled with a boost for cancer-related drugs.

SIBTMlit2¶

Run ID: SIBTMlit2
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 7/23/2018
Type: automatic
Task: abstracts
MD5: 4df01c733157aae1e7228f276e7acd90
Run description: Run 2 is based on run 1. Results are further re-ranked based on demographic information. MeSH terms are used to retrieve demographic information in the abstract. Abstracts matching the gender and age informations are favored.

SIBTMlit3¶

Run ID: SIBTMlit3
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 7/23/2018
Type: automatic
Task: abstracts
MD5: 559b66e1498371112a3df43cf0327a70
Run description: Run 3 is based on run 2. A classifier has been developed to classify papers into two categories: PM or not PM. The probability of an abstract to be PM is used to re-rank run 2.

SIBTMlit4¶

Run ID: SIBTMlit4
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 7/23/2018
Type: automatic
Task: abstracts
MD5: 6ed49fd34a3d78079ba6f58ace32a448
Run description: Run 4 is expanding the search to publications about more specific/general diseases. The simplified neoplasms hierarchy provided by NCI thesaurus is used (https://evs.nci.nih.gov/ftp1/NCI_Thesaurus/Neoplasm/Neoplasm_Core_Hierarchy.html). The run is then merged with run 3 through linear combination.

SIBTMlit5¶

Run ID: SIBTMlit5
Participant: SIBTextMining
Track: Precision Medicine
Year: 2018
Submission: 7/23/2018
Type: automatic
Task: abstracts
MD5: c24f80bff1c792a3dd4c34320d931768
Run description: Run 5 is first performing an exact run: the disease, the gene and the variant must be found in the abstract and the demographic information must be either not discussed or matching. Results from run 4 are completing the run 5.

SINAI_Base¶

Run ID: SINAI_Base
Participant: SINAI
Track: Precision Medicine
Year: 2018
Submission: 7/31/2018
Type: automatic
Task: abstracts
MD5: 8278acb287184a2078d47022e72e7e55
Run description: We create a unique query with the terms of disease, gene and demographic fields. Finally we launch the query in a Lemur IR system.

SINAI_FU¶

Run ID: SINAI_FU
Participant: SINAI
Track: Precision Medicine
Year: 2018
Submission: 7/31/2018
Type: automatic
Task: abstracts
MD5: f53b59af99137e4964d8b72edce52547
Run description: We create a query with the terms of disease, gene and demographic fields. Then we launch the query in a Lemur IR system to obtain relevant documents. Next we use MetaMap to find UMLS concepts in query and relevant documents. Finally we remove from the list of relevant documents those that do not contain any UMLS concept of the query.

SINAI_FUO¶

Run ID: SINAI_FUO
Participant: SINAI
Track: Precision Medicine
Year: 2018
Submission: 7/31/2018
Type: automatic
Task: abstracts
MD5: 9e5c6125ebfd2c648ea66b09e6aac7e7
Run description: We create a query with the terms of disease, gene and demographic fields. Then we launch the query in a Lemur IR system to obtain relevant documents. Next we use MetaMap to find UMLS concepts in query and relevant documents. We calculate the percentage of query concepts that appear in each document. Finally we remove from the list of relevant documents those that do not contain any UMLS concept of the query and reranking this list by the percentage of concepts it contains.

sq¶

Run ID: sq
Participant: Poznan
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 983b68352b7c673346c6c3aa5543187f
Run description: Simple, boolean query.

sum¶

Run ID: sum
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 468c06b520f73d146b721ebaa8cb220d
Run description: CombSUM with two language models and bm25 retreival models.

sumAbs¶

Run ID: sumAbs
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 0294fb8955b2e51c2196cb377621fca3
Run description: CombSum with two language models and bm25 retrieval model.

sumEW¶

Run ID: sumEW
Participant: Brown
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 9dfb4c963bf536b82f94bb56dabcbab5
Run description: CombSUM with wordEmbedding and no normalisation.

tinfolabBF¶

Run ID: tinfolabBF
Participant: InfoLabPM
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: trials
MD5: 9f284644b42021589ca7066e897767ac
Run description: The matching model used: BB2 Query expansion with genes weighted

tinfolabBK¶

Run ID: tinfolabBK
Participant: InfoLabPM
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: trials
MD5: 28847ceea423c270e3b3c0951271c988
Run description: The matching model used: BB2 Query expansion with genes weighted. Expanded disease term to the most preferred atom (via UMLS REST API)

tinfolabF¶

Run ID: tinfolabF
Participant: InfoLabPM
Track: Precision Medicine
Year: 2018
Submission: 8/2/2018
Type: automatic
Task: trials
MD5: 159815f2bfc842c32cefa1cecf572c4c
Run description: The matching model used: BM25 Query expansion with genes weighted

two_stage¶

Run ID: two_stage
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: e3d45439a89b20c22bc8a29950daabe1
Run description: Two stage smoothing.

two_trial1¶

Run ID: two_trial1
Participant: FDUDMIIP
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: e57f42b29f9662118fe75fed5afb52b1
Run description: two trial

UCASCT1¶

Run ID: UCASCT1
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: ae6b83f238b5ccaed186937297b4129f
Run description: This run is generated by BM25F retrieval model with Bo1 query expansion model. Some synonyms and abbreviations of diseases are added to the topic. Stopword removal and stemming are not performed when indexing and retrieving.

UCASCT2¶

Run ID: UCASCT2
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: df3dd0730f38152f4ffd099c73196635
Run description: This run is generated by BM25F retrieval model, whose weight is different from UCASCT1, with Bo1 query expansion model. Some synonyms and abbreviations of diseases are added to the topic. Stopword removal and stemming are not performed when indexing and retrieving.

UCASCT3¶

Run ID: UCASCT3
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: 726a0143492f6fd57ede4ac6f3ccabf2
Run description: This run is generated by BM25F retrieval model with Bo1 query expansion model. Some synonyms and abbreviations of diseases are added to the topic. Stopword removal and stemming are performed when indexing and retrieving.

UCASCT4¶

Run ID: UCASCT4
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: 9ff4f19abb1b8c1b325de5541ee6649d
Run description: This run is generated by BM25F retrieval model, whose weight is different from UCASCT3, with Bo1 query expansion model. Some synonyms and abbreviations of diseases are added to the topic. Stopword removal and stemming are performed when indexing and retrieving.

UCASCT5¶

Run ID: UCASCT5
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: trials
MD5: e715e47ebd9c53512c492fcd9c8198c4
Run description: This run is generated by BM25 retrieval model with KL query expansion model. Some synonyms and abbreviations of diseases are added to the topic. Stopword removal and stemming are not performed when indexing and retrieving.

UCASSA1¶

Run ID: UCASSA1
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: 3cd218fae7f8f00cf0575302607b555d
Run description: This run is generated by BM25 retrieval model with no query expansion. Some synonyms and abbreviations of diseases are added to the topic. Stopword removal and stemming are performed when indexing and retrieving.

UCASSA2¶

Run ID: UCASSA2
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: af4e2091f2c1b643c54f733a13de9418
Run description: This run is generated by BM25 retrieval model with no query expansion. Stopword removal and stemming are performed when indexing and retrieving.

UCASSA3¶

Run ID: UCASSA3
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: e9f7e04367c3d490e89d2f32f21009a6
Run description: An Initial result is generated by BM25 retrieval model with no query expansion. Some synonyms and abbreviations of diseases are added to the topic. Stopword removal and stemming are performed when indexing and retrieving. Then employ K-NRM model to re-rank the initial result.

UCASSA4¶

Run ID: UCASSA4
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: 3b725249bd258c0bb6de9e4f5c9a22a8
Run description: This run is generated by BM25 retrieval model with no query expansion. Some synonyms and abbreviations of diseases are added to the topic. Stopword removal and stemming are performed when indexing and retrieving.

UCASSA5¶

Run ID: UCASSA5
Participant: UCAS
Track: Precision Medicine
Year: 2018
Submission: 8/3/2018
Type: automatic
Task: abstracts
MD5: 72c406f9ff9384c016164b026b576b3d
Run description: An Initial result is generated by BM25 retrieval model with no query expansion. Stopword removal and stemming are performed when indexing and retrieving. Then penalize the score of articles which are not related to treatment, prevention, and prognosis, according to MeSH term.

UDInfoPMCT1¶

Run ID: UDInfoPMCT1
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 00013971d256bf4ad46ca7fed25b93c2
Run description: Term based baseline run. Use the disease and gene as it is.

UDInfoPMCT2¶

Run ID: UDInfoPMCT2
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 225c2c66acbc65d467ea62acc6c9ec32
Run description: Term base retrieval, using the gene and disease as query. Demographic information is used to filter the result

UDInfoPMCT3¶

Run ID: UDInfoPMCT3
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: aeab539545cf25aefa1f3071e0762c80
Run description: Concept based run. using original query, disease + gene as the query.

UDInfoPMCT4¶

Run ID: UDInfoPMCT4
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 6904a63c4782c0022a5c84498272a998
Run description: Concept based run. query expansion with disease ontology

UDInfoPMCT5¶

Run ID: UDInfoPMCT5
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: c691cb63d9c901fc6999ac58ba3aaf0d
Run description: Two-round. Term based run. Retrieve 5k results in first round using disease + gene + disease ontology expansion. Second round do a binary search using only the original gene field

UDInfoPMSA1¶

Run ID: UDInfoPMSA1
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: d263e22c7715ad530c4d1793e9c37a62
Run description: use the query as it is, disease + gene, term based

UDInfoPMSA2¶

Run ID: UDInfoPMSA2
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 2e55d2aba9ff0c3cda576e4e901b84c3
Run description: disease + gene, term based, query expansion using Disease ontology and genecards

UDInfoPMSA3¶

Run ID: UDInfoPMSA3
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 41fc44aabc7662b542d5a70e8984cbaf
Run description: Concept based run. using original query, disease + gene as the query.

UDInfoPMSA4¶

Run ID: UDInfoPMSA4
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: c7447aa0f34a0b1b1c77e0e0b243b351
Run description: Concept based run. query expansion with disease ontology

UDInfoPMSA5¶

Run ID: UDInfoPMSA5
Participant: udel_fang
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 8680fb2d91b5174cb3295cbb205650bd
Run description: Two-round. Term based run. Retrieve 5k results in first round using disease + gene + disease ontology expansion. Second round do a binary search using only the original gene field

UNTIIA_DGES¶

Run ID: UNTIIA_DGES
Participant: UNTIIA
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: 4833c86f77dab05f5ef310c3f8efaa2d
Run description: This run was performed using queries constructed with synonyms of diseases in the topics acquired from MetaMap and synonyms of genes in the topic acquired from NCBI genBank as query expansion with Solr logical query language. The reranking was based on regression analysis, which is a supervised learning algorithm. The training data of the model is from TREC 2017 PM track.

UNTIIA_DGEWS¶

Run ID: UNTIIA_DGEWS
Participant: UNTIIA
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: 397056c2b3c28a513e4f1204b51da465
Run description: This run was performed using queries constructed using disease weighted terms and their synonyms, gene weighted terms and their synonyms with Solr logical query language. The reranking was based on regression analysis, which is a supervised learning algorithm. The training data of the model is from TREC 2017 PM track.

UNTIIA_DGEWU¶

Run ID: UNTIIA_DGEWU
Participant: UNTIIA
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: eb0d97e1d20142533dfed0716570a76a
Run description: This run was performed using queries constructed using disease weighted terms and their synonyms, gene weighted terms and their synonyms with Solr logical query language. The reranking was based on Doc2Vec, a unsupervised learning technique based on Word2Vec.

UNTIIA_DGS¶

Run ID: UNTIIA_DGS
Participant: UNTIIA
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: d0ad53e245a5a2ebdd782227cc889c0d
Run description: This run was performed using queries constructed with diseases and genes from the given topics as baseline method. The reranking was based on regression analysis, which is a supervised learning algorithm. The training data of the model is from TREC 2017 PM track.

UNTIIA_WTU¶

Run ID: UNTIIA_WTU
Participant: UNTIIA
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: 5dc15c17ebbc60f0f8e2c954bcda115e
Run description: This run was performed using queries constructed using disease weighted terms and their synonyms, gene weighted terms and their synonyms, disease specific treatment terms acquired from American cancer society with Solr logical query language. The reranking was based on Doc2Vec, a unsupervised learning technique based on Word2Vec.

UTDHLTRI_NL¶

Run ID: UTDHLTRI_NL
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: cfedde998b594a07a6376d7d1b942236
Run description: This run does retrieval using methods use in the previous year with a few adaptations for the new topic format. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), and Snomed were used.

UTDHLTRI_NLT¶

Run ID: UTDHLTRI_NLT
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 8eede80164453b405072dac1f79f9774
Run description: This run does retrieval using methods use in the previous year with a few adaptations for the new topic format. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), and Snomed were used.

UTDHLTRI_RA¶

Run ID: UTDHLTRI_RA
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: ab57b34b13ed4729ce7df5eca6600d16
Run description: This run does an initial retrieval using methods use in the previous year, and then applies a learning-to-rank algorithm to re-rank the results. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), Snomed, and Ranklib were used.

UTDHLTRI_RAT¶

Run ID: UTDHLTRI_RAT
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 001278f97be64fd88f9d84cbec2c1e5b
Run description: This run does an initial retrieval using methods use in the previous year, and then applies a learning-to-rank algorithm to re-rank the results. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), Snomed, and Ranklib were used.

UTDHLTRI_RF¶

Run ID: UTDHLTRI_RF
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 1307e9adf9dc00d5bfd141a31b307f67
Run description: This run does an initial retrieval using methods use in the previous year, and then applies a learning-to-rank algorithm to re-rank the results. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), Snomed, and Ranklib were used.

UTDHLTRI_RFT¶

Run ID: UTDHLTRI_RFT
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: d62ca3cfd285750e85321064f2ccad28
Run description: This run does an initial retrieval using methods use in the previous year, and then applies a learning-to-rank algorithm to re-rank the results. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), Snomed, and Ranklib were used.

UTDHLTRI_SF¶

Run ID: UTDHLTRI_SF
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: e93a916cb89468a13841c00013d4acf9
Run description: This run does an initial retrieval using methods use in the previous year, and then applies a learning-to-rank algorithm to re-rank the results. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), Snomed, and Ranklib were used.

UTDHLTRI_SFT¶

Run ID: UTDHLTRI_SFT
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: fd5cfdb9a370b0bbbe53b7c77e6e1381
Run description: This run does an initial retrieval using methods use in the previous year, and then applies a learning-to-rank algorithm to re-rank the results. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), Snomed, and Ranklib were used.

UTDHLTRI_SS¶

Run ID: UTDHLTRI_SS
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 60df297116dc2130cab642dfa1cb469f
Run description: This run does an initial retrieval using methods use in the previous year, and then applies a learning-to-rank algorithm to re-rank the results. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), Snomed, and Ranklib were used.

UTDHLTRI_SST¶

Run ID: UTDHLTRI_SST
Participant: UTDHLTRI
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: trials
MD5: 169955678823f24a91e784f52fbdd4b1
Run description: This run does an initial retrieval using methods use in the previous year, and then applies a learning-to-rank algorithm to re-rank the results. UMLS, Catalogue of Somatic Mutations in Cancer (COSMIC), FDA, Drug-Gene Interaction Database (DGIdb), Snomed, and Ranklib were used.

UVAEXPBOOST¶

Run ID: UVAEXPBOOST
Participant: UVA_ART
Track: Precision Medicine
Year: 2018
Submission: 8/4/2018
Type: automatic
Task: abstracts
MD5: ccef2e76b64cd3362f55a914d2fe483e
Run description: Resources used: cTAKES (for annotating documents) UMLS (for constructing concept graph) Elasticsearch (for query engine) A subset of documents matching a rough text filter were annotated using cTAKES to produce a set of identified concepts for each document. The documents and the concept terms were indexed in Elasticsearch. We built a concept graph from several UMLS ontologies and implemented a relatedness metric to quantify relationships between concepts. This was used to expand the topic data to a wider group of concepts. The expanded topics were input into a query format for Elasticsearch that emphasized CUI matches (especially disease and gene) and boosted results including terms that refer to treatment and prognosis, and negatively boosted some terms that refer to non-cancer or non-human terms.

UVAEXPBSTDIF¶

Run ID: UVAEXPBSTDIF
Participant: UVA_ART
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: 699686de35aadf64b0d233cd2432995d
Run description: Resources used: cTAKES (for annotating documents) UMLS (for constructing concept graph) Elasticsearch (for query engine) A subset of documents matching a rough text filter were annotated using cTAKES to produce a set of identified concepts for each document. The documents and the concept terms were indexed in Elasticsearch. We built a concept graph from several UMLS ontologies and implemented a relatedness metric to quantify relationships between concepts. This was used to expand the topic data to a wider group of concepts. The expanded topics were input into a query format for Elasticsearch that emphasized CUI matches (especially disease and gene) and boosted results including terms that refer to treatment and prognosis, and negatively boosted some terms that refer to non-cancer or non-human terms. This run used a different boosting ratio than UVAEXPBOOST.

UVAEXPBSTEXT¶

Run ID: UVAEXPBSTEXT
Participant: UVA_ART
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: dd99815234ecb5123a2d80adb464e7c0
Run description: Resources used: cTAKES (for annotating documents) UMLS (for constructing concept graph) Elasticsearch (for query engine) A subset of documents matching a rough text filter were annotated using cTAKES to produce a set of identified concepts for each document. The documents and the concept terms were indexed in Elasticsearch. We built a concept graph from several UMLS ontologies and implemented a relatedness metric to quantify relationships between concepts. This was used to expand the topic data to a wider group of concepts. The expanded topics were input into a query format for Elasticsearch that emphasized CUI matches (especially disease and gene) and boosted results including terms that refer to treatment and prognosis, and negatively boosted some terms that refer to non-cancer or non-human terms. This is similar to UVAEXPBOOST but includes extra CUI terms for clinical trials and related to apoptosis and remission, which are boosted, and terms for screening and detection which are negatively boosted.

UVAEXPBSTNEG¶

Run ID: UVAEXPBSTNEG
Participant: UVA_ART
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: f78d7f48bb9313bae873bc718a3d17ce
Run description: Resources used: cTAKES (for annotating documents) UMLS (for constructing concept graph) Elasticsearch (for query engine) A subset of documents matching a rough text filter were annotated using cTAKES to produce a set of identified concepts for each document. The documents and the concept terms were indexed in Elasticsearch. We built a concept graph from several UMLS ontologies and implemented a relatedness metric to quantify relationships between concepts. This was used to expand the topic data to a wider group of concepts. The expanded topics were input into a query format for Elasticsearch that emphasized CUI matches (especially disease and gene) and boosted results including CUIs that refer to treatment and clinical trial concepts, and negatively boosted some CUIs that refer to cancer screening and detection.

UVAEXPBSTSHD¶

Run ID: UVAEXPBSTSHD
Participant: UVA_ART
Track: Precision Medicine
Year: 2018
Submission: 8/5/2018
Type: automatic
Task: abstracts
MD5: aa91ec0cceef479ab4bb714b2aa8eb20
Run description: Resources used: cTAKES (for annotating documents) UMLS (for constructing concept graph) Elasticsearch (for query engine) A subset of documents matching a rough text filter were annotated using cTAKES to produce a set of identified concepts for each document. The documents and the concept terms were indexed in Elasticsearch. We built a concept graph from several UMLS ontologies and implemented a relatedness metric to quantify relationships between concepts. This was used to expand the topic data to a wider group of concepts. The expanded topics were input into a query format for Elasticsearch that emphasized CUI matches (especially disease and gene) and boosted results including terms that refer to treatment and prognosis, and negatively boosted some terms that refer to non-cancer or non-human terms. This run differs in that it leaves gene as a "should" field rather than a "must" field as with other runs. It includes the extra terms from UVAEXPBSTNEG.