You are sorting a reading list before a genetic evidence review.

The question that will be answered from this list is whether variants in
{gene_name} cause or are associated with {phenotype}.

That phenotype is written under other names in the literature, and an article
using any of them is writing about the same phenotype:

{synonym_note}

Below are the candidate articles:

{articles_list}

Put every article listed above in one of four bands. Return one entry per
article, none omitted and none invented.

  4  Strong human evidence. It reports variants in {gene_name} in people who
     have or were measured for {phenotype}: probands, families, cohorts, or
     population data. A study arguing against the association on human data
     belongs here too, since that is evidence about the same question.
  3  Weak human evidence. It concerns {gene_name} in people, but the link to
     {phenotype} falls short: a clinically related condition, a phenotype
     described without the variants being shown to cause it, a single
     uncertain observation, or a review of human work reporting no new cases.
  2  Evidence that is not human. Animal models, cells, proteins, structures,
     biochemistry, or the function of {gene_name} in the tissue or pathway
     {phenotype} involves.
  1  No bearing on the question at all. Use this above all when the letters
     {gene_name} in the article are not this gene: a different gene whose
     symbol merely starts the same way, a person's name inside a disease
     eponym, a laboratory abbreviation, or an ordinary word of English.

Only a full reading of the article decides these things, and you have an
abstract. So when two bands are open to you, take the higher one. A band of 1
in particular is a claim that the article cannot possibly matter, and it should
be made only when you are sure; an article whose subject you cannot make out
from its abstract is a 2, not a 1. Understating one article costs the review a
piece of evidence it will never see, while overstating one costs it a minute of
reading.

Judge each article on its own. Do not rank them against each other, do not
decide how convincing the evidence is, and do not try to shorten the list: a
list where every article is a 4 is a valid answer, and so is one where every
article is a 1.

Return your answer in this exact XML format:
<response>
<article><pmid>12345678</pmid><band>4</band></article>
<article><pmid>23456789</pmid><band>1</band></article>
...
</response>
