Full Text for Variant Curation: Read Before You Buy the PDF

Variant curation is the work of deciding whether a DNA change contributes to disease. A genetic test finds the change; a curator weighs the evidence to classify it as disease-causing, harmless or uncertain. That often means reading reports about other patients, with the decisive details in the full text or supplements.

A DeepDyve subscription lets teams read papers in its LitStream collection in the browser without a per-article charge, then buy only the PDFs they need to retain. Curators score evidence from the paper instead of guessing from the abstract.

Why curators need more than an abstract

Clinical labs use the ACMG/AMP variant-classification framework, with refinements from ClinGen. Its evidence codes label findings that support a classification. Many require details that a database entry or abstract may not provide:

• PS4, affected individuals. In some rare-disease applications, curators count unrelated patients with the variant. They check symptoms and avoid counting the same patient across multiple papers.
• PP1, segregation. Does the variant track with disease through a family? Family trees and relatives’ test results help establish the strength of this evidence.
• PM3, in trans. For recessive disorders, curators check whether the variant is on the other copy of the gene from a disease-causing variant, often using parental testing.
• PS2 and PM6, de novo. A variant appears in a child but not the parents. Its evidential weight depends on parentage confirmation, the patient’s symptoms and the number of observations.
• PS3, functional studies. Experiments may show a damaging effect, but curators must assess the assay’s controls and validation in the methods and supplements.
• PM5, another change at the same residue. A different disease-causing amino-acid substitution at the same position may support classification. Curators check the evidence behind that earlier classification.

“The variant segregated with disease” is not enough to judge the strength of the evidence. The curator needs the family data behind the statement. These are examples, not scoring instructions: the applicable gene- and disease-specific guidance matters.

Why getting the full text is hard

Older reports still matter. An early case report may remain important even decades later. A journal’s move to open access does not necessarily make its older papers free.

Reading does not always yield usable evidence. Candidate papers may describe a different variant, a previously reported family or no relevant patient data. Paying to open each one adds up, whether or not it supports the classification.

Library access can be patchy. Commercial labs and small clinical teams may lack subscriptions, while relevant reports span genetics and disease-specific journals from several publishers.

Where DeepDyve fits

1. Read before buying a PDF. LitHub by DeepDyve provides subscription-based browser access to the LitStream collection. Reading is included; downloadable PDFs are separate purchases.

LitStream includes papers from genetics journals such as Human Mutation, Clinical Genetics, Human Genetics and European Journal of Human Genetics. Coverage varies by journal and publication year; check the specific articles and older volumes your team needs.

2. Find passages with AI, then verify them. The AI Research Assistant answers questions about eligible papers and points to source passages. Ask whether parentage was confirmed or which relatives were tested, then check the text yourself. Evidence scoring remains the curator’s responsibility. Do not enter patient identifiers in prompts.

3. Organize the supporting literature. Shared folders and collaborative annotations let teams group papers by gene and variant, record the evidence codes they support and mark relevant passages. Tags help individual curators organize their reading but are not shared with the team. Your curation system remains the authoritative record of the classification and its rationale. Saving a reference does not give permanent access to a streamed paper.

4. Buy only what you need to retain. Available partner-publisher PDFs are sold at the publisher’s price without a DeepDyve markup. Enterprise users can order other available articles through Copyright Clearance Center (CCC), which adds a service fee. Check the license for your intended storage and sharing.

5. Avoid duplicate purchases and manage spending. Enterprise surfaces copies already held in the team’s Research Archive. Administrators can set monthly purchase limits.

What to confirm before relying on it

• Journal and archive coverage. Check recent reference lists against the collection at the article level; a listed journal does not guarantee access to every issue. Elsevier is not in LitStream; gaps include The American Journal of Human Genetics. Check CCC ordering for papers outside the collection; do not assume they can be streamed before purchase.
• Supplementary files. Family trees and genotype tables may be hosted separately on publishers’ sites. Test access for your core journals.
• AI eligibility and data handling. Not every readable paper supports AI analysis. Test representative papers and ask where prompts are processed and how long they are retained.
• Existing library access. If your institution already supplies the papers and older volumes you need, another subscription may add little.

How to test it in a month

1. Measure coverage. List papers cited in last month’s completed reports and count how many are readable in LitStream.
2. Measure avoidable purchases. Total what you paid for papers you read but did not cite. Check how many were streamable.
3. Compare access time. For one variant, time getting the same candidate papers through DeepDyve and your current route.
4. Test evidence retrieval. Organize and annotate papers for three variants. Ask a second curator to locate the supporting passages and compare them with the recorded rationale.
5. Test duplicate detection. Use a paper already purchased for the archive and confirm another curator sees the existing copy.

FAQ

Does DeepDyve replace ClinVar, LitVar, Mastermind or HGMD?
No. Those tools help curators find variants, reported classifications or relevant literature. DeepDyve provides access to papers within its coverage, not a substitute for curation tools.

Can the AI Research Assistant classify a variant?
No. It helps locate information in eligible papers. Curators must verify the evidence and apply the classification criteria.

Can we use the Personal plan for diagnostic work?
No. Personal is not licensed for business use. Curation teams need Enterprise.

Does a subscription include permanent copies?
Not of streamed papers. Browser access is subscription-based; PDF purchases are separate. Check retention and sharing rights for any copies you buy.

If your curation team is still getting full text one purchase at a time, talk to a DeepDyve client specialist about Enterprise and check one month of your lab’s reference lists against the collection.

DeepDyve pricing
DeepDyve features
Browse journals on DeepDyve